Elevator dynamics model correction system and correction method

Abstract

The present invention relates to the field of dynamic model technology, and in particular to an elevator dynamic model correction system and correction method. The method comprises the following steps: obtaining basic data of the elevator system and historical operation data of the elevator; performing multi-scale dynamic characteristic analysis based on the basic data of the elevator system to generate multi-scale dynamic characteristic data; performing multi-scale stability analysis on the multi-scale dynamic characteristic data based on the historical operation data of the elevator, thereby obtaining multi-scale stability data; performing subsystem coupling sensitivity analysis on the multi-scale stability data based on the structural parameters in the basic data of the elevator system, thereby generating multi-scale coupling sensitivity data; performing global nonlinear characteristic mapping processing based on the multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model. The present invention can estimate the state that is difficult to measure directly by constructing an extended state observer based on a high-frequency vibration mode, thereby improving the comprehensiveness and accuracy of system monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamics models, and in particular to an elevator dynamics model correction system and correction method. Background Art

[0002] The elevator dynamics model is a mathematical expression that describes the motion behavior of the elevator system. It covers various physical phenomena and interactions during the operation of the elevator. The main components of the elevator dynamics model are the mass system (car mass, counterweight mass and wire rope mass), elastic elements (wire rope elasticity and guide rail elasticity), friction (friction between the car and the guide rail, friction of the pulley system and friction of the mechanical transmission system), drive system (motor characteristics and transmission mechanism), control system (speed controller and position controller), external factors (wind load and earthquake impact), nonlinear effects (nonlinear stiffness of wire rope and nonlinear characteristics of friction), energy loss (air resistance and various friction losses) and dynamic equations.

[0003] There are often the following technical difficulties when correcting elevator dynamic models: the elevator system has many nonlinear and time-varying characteristics, which makes model correction complicated, such as the change of wire rope stiffness with position, the change of friction with speed and temperature, and the change of load mass with the number of passengers; the elevator system consists of multiple subsystems such as the car, counterweight, wire rope, etc., and the interactions between them are complex. It is a major challenge to accurately describe the coupling relationship between these subsystems; high-frequency vibrations may be generated during elevator operation. These vibration modes are difficult to accurately capture in low-order models, but have an important impact on passenger comfort; the elevator system contains fast (such as vibration) and slow (such as thermal effects) dynamic processes that span multiple time scales, such as high-frequency vibrations of wire ropes and low-frequency motions of elevators, short-term control dynamics and long-term wear effects. Summary of the invention

[0004] Based on this, it is necessary for the present invention to provide an elevator dynamics model correction system and correction method to solve at least one of the above technical problems.

[0005] To achieve the above object, a method for correcting an elevator dynamics model comprises the following steps:

[0006] Step S1: acquiring basic data of the elevator system and historical operation data of the elevator; performing multi-scale dynamic characteristic analysis according to the basic data of the elevator system to generate multi-scale dynamic characteristic data; performing multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator to obtain multi-scale stability data;

[0007] Step S2: performing subsystem coupling sensitivity analysis on the multi-scale stability data according to the structural parameters in the elevator system basic data, thereby generating multi-scale coupling sensitivity data;

[0008] Step S3: Perform global nonlinear characteristic mapping processing on the multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; perform parameter identification strategy processing on the global nonlinear dynamic model to generate an adaptive parameter identification strategy; perform time-varying characteristic identification processing on the multi-scale stability data based on the historical operation data of the elevator to generate a time-varying parameter update strategy;

[0009] Step S4: acquiring real-time status monitoring data of the elevator system, identifying abnormal status based on the real-time status monitoring data, and obtaining abnormal warning signal data; performing multi-time scale dynamic correction processing on the abnormal warning signal data using an adaptive parameter identification strategy and a time-varying parameter update strategy, and obtaining dynamic correction parameter data;

[0010] Step S5: performing model update according to the dynamic correction parameter data, and performing model accuracy evaluation to generate model correction effect data; re-optimizing the model structure according to the adaptive parameter identification strategy based on the model correction effect data to generate optimized model structure data;

[0011] Step S6: construct an extended state observer based on the high-frequency vibration mode according to the optimization model structure data, and use the extended state observer to estimate the state that is difficult to measure directly, so as to obtain high-frequency vibration estimation data; perform multi-subsystem collaborative optimization according to the high-frequency vibration estimation data and multi-scale coupling sensitivity data, and generate inter-subsystem coupling optimization parameters; correct the global nonlinear dynamic model according to the inter-subsystem coupling optimization parameters to obtain a corrected elevator dynamic model.

[0012] The present invention obtains basic data and historical operation data, performs multi-scale dynamic characteristic analysis and stability analysis, and can provide an in-depth understanding of the dynamic characteristics of the elevator system; this helps to identify the performance of the system at different time scales, laying the foundation for subsequent analysis and optimization; it provides comprehensive dynamic characteristics of the elevator system under different operating conditions, which helps to discover potential stability problems. The structural parameters are combined with stability data to perform subsystem coupling sensitivity analysis, which can identify which parts of the system have the greatest impact on the overall stability; it clarifies the key components and parameters in the system, and provides a direction for targeted optimization. The generation of global nonlinear characteristic mapping and parameter identification strategies helps to establish a more accurate elevator dynamic model, and can adaptively adjust model parameters to adapt to different operating conditions; it improves the adaptability and prediction accuracy of the model, and provides a powerful tool for real-time monitoring and control. Real-time status monitoring and abnormal state identification can timely discover and warn possible problems, while dynamic correction processing can reduce false alarms and missed alarms; it enhances the safety and reliability of the system, and through real-time monitoring and correction, it can prevent the occurrence of faults and accidents. Model updating and accuracy assessment ensure that the model always reflects the latest system status, while model structure re-optimization further improves the performance of the model; through continuous model updating and optimization, the accuracy and effectiveness of the model are maintained, and the overall performance of the system is improved. Building an extended state observer and performing multi-subsystem collaborative optimization can more accurately estimate difficult-to-measure states and optimize the coupling between subsystems; improve the accuracy of system state estimation, and by optimizing coupling parameters, improve the overall collaborative work effect and operating efficiency of the system. In general, this series of steps significantly improves the stability, safety and efficiency of the elevator system through in-depth analysis, real-time monitoring, dynamic correction and model optimization. In this way, the complex dynamic behavior of the elevator system can be better understood and controlled, thereby achieving better system performance and longer service life.

[0013] The present invention also provides an elevator dynamics model correction system, which is used to execute the above-mentioned elevator dynamics model correction method. The elevator dynamics model correction system comprises:

[0014] The basic data and multi-scale analysis module is used to obtain the basic data of the elevator system and the historical operation data of the elevator; perform multi-scale dynamic characteristics analysis based on the basic data of the elevator system to generate multi-scale dynamic characteristics data; perform multi-scale stability analysis on the multi-scale dynamic characteristics data based on the historical operation data of the elevator to obtain multi-scale stability data;

[0015] A coupling sensitivity analysis module is used to perform subsystem coupling sensitivity analysis on multi-scale stability data according to structural parameters in the basic data of the elevator system, thereby generating multi-scale coupling sensitivity data;

[0016] The nonlinear model and parameter identification module is used to perform global nonlinear characteristic mapping processing based on multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; perform parameter identification strategy processing based on the global nonlinear dynamic model to generate an adaptive parameter identification strategy; perform time-varying characteristic identification processing on multi-scale stability data based on the historical operation data of the elevator to generate a time-varying parameter update strategy;

[0017] The real-time monitoring and dynamic correction module is used to obtain real-time status monitoring data of the elevator system, identify abnormal status based on the real-time status monitoring data, and obtain abnormal warning signal data; use adaptive parameter identification strategy and time-varying parameter update strategy to perform multi-time scale dynamic correction processing on abnormal warning signal data to obtain dynamic correction parameter data;

[0018] The model updating and optimization module is used to perform model updating according to the dynamic correction parameter data, evaluate the model accuracy, and generate model correction effect data; based on the model correction effect data, the model structure is re-optimized according to the adaptive parameter identification strategy to generate optimized model structure data;

[0019] The extended observation and correction module is used to construct an extended state observer based on the high-frequency vibration mode according to the optimization model structure data, and use the extended state observer to estimate the state that is difficult to measure directly, so as to obtain high-frequency vibration estimation data; perform multi-subsystem collaborative optimization based on the high-frequency vibration estimation data and multi-scale coupling sensitivity data to generate inter-subsystem coupling optimization parameters; and correct the global nonlinear dynamic model according to the inter-subsystem coupling optimization parameters to obtain the corrected elevator dynamic model.

[0020] The present invention integrates basic data and historical operation data to conduct multi-scale dynamics and stability analysis; it provides a basis for in-depth understanding of the operating status of the elevator system, can identify the behavioral characteristics and stability status of the system at different time scales, and lay a solid foundation for subsequent analysis and optimization. Based on structural parameters and stability data, the coupling sensitivity between subsystems is analyzed; the influence of the interaction between different subsystems on the overall stability of the system is revealed, and the key factors affecting the performance of the system are identified, which provides a basis for the fine adjustment and optimization of the system. It processes global nonlinear characteristics, constructs dynamic models, and develops parameter identification and time-varying characteristic identification strategies; by constructing accurate nonlinear dynamic models and adaptive parameter identification strategies, the model's ability to predict system behavior is improved, and it can adapt to changes in system operation, enhancing the practicality and flexibility of the model. Real-time monitoring of system status, identification of anomalies, and dynamic correction of multiple time scales are performed; through real-time monitoring and anomaly warning, potential problems can be discovered and responded to in a timely manner, and the dynamic correction strategy ensures that the system can maintain good performance under various working conditions, improving the reliability and safety of the system. Update the model according to the correction parameters, evaluate the model accuracy, and perform structural optimization; the continuous updating and optimization of the model ensures the accuracy and adaptability of the model, making the design and operation of the elevator system more in line with actual needs and improving the efficiency and performance of the system. Build an extended state observer to estimate the difficult-to-measure state, perform multi-subsystem collaborative optimization and dynamic model correction; through high-frequency vibration estimation and coupling optimization between subsystems, the understanding and control capabilities of the complex dynamic behavior of the system are further improved. The corrected model is more accurate and can better guide the design and maintenance of the elevator system. Overall, this process forms a closed-loop system optimization and control process from data collection to model construction, parameter identification, real-time monitoring, dynamic correction, and then to model update and optimization. The output of each module provides input for the next module, ensuring the consistency and effectiveness of the entire system analysis and optimization; through this systematic approach, the performance, reliability and safety of the elevator system can be significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0022] Figure 1 It is a schematic diagram of the steps of the elevator dynamics model correction method of the present invention;

[0023] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0024] Figure 3 for Figure 1Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0025] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0027] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0028] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for correcting an elevator dynamics model, the method comprising the following steps:

[0029] Step S1: acquiring basic data of the elevator system and historical operation data of the elevator; performing multi-scale dynamic characteristic analysis according to the basic data of the elevator system to generate multi-scale dynamic characteristic data; performing multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator to obtain multi-scale stability data;

[0030] Step S2: performing subsystem coupling sensitivity analysis on the multi-scale stability data according to the structural parameters in the elevator system basic data, thereby generating multi-scale coupling sensitivity data;

[0031] Step S3: Perform global nonlinear characteristic mapping processing on the multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; perform parameter identification strategy processing on the global nonlinear dynamic model to generate an adaptive parameter identification strategy; perform time-varying characteristic identification processing on the multi-scale stability data based on the historical operation data of the elevator to generate a time-varying parameter update strategy;

[0032] Step S4: acquiring real-time status monitoring data of the elevator system, identifying abnormal status based on the real-time status monitoring data, and obtaining abnormal warning signal data; performing multi-time scale dynamic correction processing on the abnormal warning signal data using an adaptive parameter identification strategy and a time-varying parameter update strategy, and obtaining dynamic correction parameter data;

[0033] Step S5: performing model update according to the dynamic correction parameter data, and performing model accuracy evaluation to generate model correction effect data; re-optimizing the model structure according to the adaptive parameter identification strategy based on the model correction effect data to generate optimized model structure data;

[0034] Step S6: construct an extended state observer based on the high-frequency vibration mode according to the optimization model structure data, and use the extended state observer to estimate the state that is difficult to measure directly, so as to obtain high-frequency vibration estimation data; perform multi-subsystem collaborative optimization according to the high-frequency vibration estimation data and multi-scale coupling sensitivity data, and generate inter-subsystem coupling optimization parameters; correct the global nonlinear dynamic model according to the inter-subsystem coupling optimization parameters to obtain a corrected elevator dynamic model.

[0035] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a method for correcting an elevator dynamics model of the present invention. In this example, the topological structure visualization method of the refrigerant gas sensor includes the following steps:

[0036] Step S1: acquiring basic data of the elevator system and historical operation data of the elevator; performing multi-scale dynamic characteristic analysis according to the basic data of the elevator system to generate multi-scale dynamic characteristic data; performing multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator to obtain multi-scale stability data;

[0037] The embodiment of the present invention uses the elevator control system and the sensor network to obtain the basic data of the elevator system, which includes the elevator car mass, counterweight mass, wire rope parameters, motor and transmission system parameters, guide rail guide shoe geometry and material parameters, and elevator shaft environmental parameters. At the same time, the historical operation data is collected through the elevator's Internet of Things platform, covering information such as position, speed, acceleration, electrical parameters, vibration level, load change, etc. Then, a multi-scale dynamic analysis method, such as wavelet transform or empirical mode decomposition (EMD), is used to process the basic data, and the dynamic characteristics of the elevator are decomposed into three time scales: high-frequency vibration, medium-frequency motion, and low-frequency trend. The dynamic characteristic data corresponding to each time scale are respectively reduced in dimension and identified by feature extraction algorithms, such as principal component analysis (PCA) or independent component analysis (ICA), to generate corresponding multi-scale dynamic characteristic data. Subsequently, these multi-scale characteristics are analyzed for stability in combination with the historical operation data, and the Lyapunov stability theory and eigenvalue analysis method are used to calculate the system stability boundary and key stability parameters at each time scale to obtain the final multi-scale stability data.

[0038] Step S2: performing subsystem coupling sensitivity analysis on the multi-scale stability data according to the structural parameters in the elevator system basic data, thereby generating multi-scale coupling sensitivity data;

[0039] The embodiment of the present invention constructs a multi-subsystem model of an elevator system based on the structural parameters of the elevator, including a car subsystem, a counterweight subsystem, a wire rope subsystem, a drive subsystem and a guide subsystem. Subsequently, the coupling relationship matrix between the subsystems is established by using the multi-scale stability data and the multi-subsystem model through a global sensitivity analysis tool (such as the Sobol sensitivity analysis method) to reflect the degree of mutual influence between different subsystems. Next, the orthogonal experimental design method is adopted to design a parameter variation experiment to analyze the influence of key coupling parameters on system stability, and the support vector regression (SVR) algorithm is used to establish a mapping relationship model between subsystem coupling parameters and system stability. Finally, for each subsystem coupling parameter, its sensitivity coefficient to system stability is calculated, and the importance of the parameters is ranked according to the sensitivity coefficient to generate a list of key coupling parameters; based on the list, a multi-scale coupling sensitivity model reflecting the influence of subsystem coupling on system stability at different time scales is constructed, and the model is analyzed and optimized by using Monte Carlo simulation or Bayesian optimization method, and finally multi-scale coupling sensitivity data is generated.

[0040] Step S3: Perform global nonlinear characteristic mapping processing on the multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; perform parameter identification strategy processing on the global nonlinear dynamic model to generate an adaptive parameter identification strategy; perform time-varying characteristic identification processing on the multi-scale stability data based on the historical operation data of the elevator to generate a time-varying parameter update strategy;

[0041] The embodiment of the present invention first uses a clustering algorithm (such as K-means or DBSCAN) to analyze the coupling sensitivity data, identify the main nonlinear characteristics in the elevator system, and generate nonlinear characteristic classification data according to the classification results. Next, a nonlinear regression model (such as polynomial regression or neural network) is used to fit the nonlinear characteristics and construct a nonlinear mapping function set. Then, according to the preset basic dynamic equation of the elevator system, these nonlinear mapping function sets are integrated to form a global nonlinear dynamic model. On this basis, a parameter identification model based on a machine learning algorithm (such as stochastic gradient descent SGD) is designed, and an adaptive learning rate strategy (such as Adam optimizer) is used to train and optimize the model, thereby generating an adaptive parameter identification strategy. At the same time, a time series analysis is performed on the historical operation data of the elevator, and a sliding time window method is used to identify the time-varying characteristics of the system parameters, and a time-varying parameter estimation algorithm is designed based on a Kalman filter algorithm (such as an extended Kalman filter EKF), and an update strategy is generated for the parameter time-varying characteristic data, and finally a time-varying parameter update strategy is formed.

[0042] Step S4: acquiring real-time status monitoring data of the elevator system, identifying abnormal status based on the real-time status monitoring data, and obtaining abnormal warning signal data; performing multi-time scale dynamic correction processing on the abnormal warning signal data using an adaptive parameter identification strategy and a time-varying parameter update strategy, and obtaining dynamic correction parameter data;

[0043] After the embodiment of the present invention has established a global nonlinear dynamic model and obtained an adaptive parameter identification strategy and a time-varying parameter update strategy, the next step is to perform dynamic correction through a real-time monitoring system. First, a sensor network is deployed in the elevator system to obtain the system's operating status data in real time, including displacement, velocity, acceleration and other information. The acquired real-time status monitoring data is preprocessed, including denoising, filtering and data standardization to ensure data quality. Then, a machine learning model (such as a support vector machine SVM or a random forest) is used to establish a normal operating mode and an abnormal state feature library, and the real-time state data is compared with the feature library to identify potential abnormal states and generate abnormal warning signal data. For these abnormal signals, combined with an adaptive parameter identification strategy and a time-varying parameter update strategy, a multi-time scale dynamic correction method (such as a sliding average method for short-term dynamic correction and an exponential smoothing method for long-term trend correction) is used for processing to obtain comprehensive dynamic correction parameter data.

[0044] Step S5: performing model update according to the dynamic correction parameter data, and performing model accuracy evaluation to generate model correction effect data; re-optimizing the model structure according to the adaptive parameter identification strategy based on the model correction effect data to generate optimized model structure data;

[0045] After completing real-time monitoring and dynamic correction, the embodiment of the present invention uses the obtained dynamic correction parameter data to update the global nonlinear dynamic model. First, the dynamic correction parameters are applied to the global model to adjust the parameters and structure of the model, thereby generating an updated dynamic model. Then, the model is used to predict the short-term behavior of the elevator system, and the prediction results are compared with the real-time monitoring data. The accuracy of the model is evaluated by calculating error indicators (such as mean square error MSE, mean absolute error MAE, etc.). According to the evaluation results, the effectiveness of the model correction is evaluated and the model correction effect data is generated. Finally, according to the correction effect data, the learning rate and forgetting factor in the adaptive parameter identification strategy are adjusted to generate a parameter identification optimization strategy, and the structure of the dynamic model is further optimized to obtain the final optimized model structure data.

[0046] Step S6: construct an extended state observer based on the high-frequency vibration mode according to the optimization model structure data, and use the extended state observer to estimate the state that is difficult to measure directly, so as to obtain high-frequency vibration estimation data; perform multi-subsystem collaborative optimization according to the high-frequency vibration estimation data and multi-scale coupling sensitivity data, and generate inter-subsystem coupling optimization parameters; correct the global nonlinear dynamic model according to the inter-subsystem coupling optimization parameters to obtain a corrected elevator dynamic model.

[0047] After obtaining the optimized dynamic model structure data, the embodiment of the present invention further performs high-frequency vibration estimation and subsystem collaborative optimization. First, the multi-scale dynamic characteristic data is analyzed using the empirical mode decomposition method to identify the high-frequency vibration mode in the elevator system and obtain high-frequency vibration mode recognition data. Then, based on the optimized model structure data, the high-frequency vibration mode recognition data is integrated into the state space model, an extended state space model is constructed, and an extended state observer is designed to determine the observer's gain matrix and the strategy for processing nonlinear terms. Next, the extended state observer is used to estimate the state that is difficult to measure directly to obtain high-frequency vibration estimation data. Finally, the high-frequency vibration estimation data and multi-scale coupling sensitivity data are combined to perform collaborative optimization of multiple subsystems in the elevator system, generate coupling optimization parameters between subsystems, and finally correct the global nonlinear dynamic model based on these parameters to obtain a corrected elevator dynamic model.

[0048] Preferably, step S1 comprises the following steps:

[0049] Step S11: acquiring basic data of the elevator system, including structural parameters based on car mass, counterweight mass and wire rope parameters, motor and transmission system parameters, geometry and material parameters of guide rails and guide shoes, and elevator shaft environment parameters;

[0050] Step S12: collecting the position, speed, acceleration, electrical parameters, vibration level and load changes during the operation of the elevator, so as to obtain the historical operation data of the elevator;

[0051] Step S13: performing empirical mode decomposition on the elevator historical operation data and decomposing it into components of different time scales, thereby obtaining a multi-scale decomposition data set, wherein the different time scales include high-frequency vibration, medium-frequency motion, and low-frequency trend;

[0052] Step S14: establishing an initial theoretical dynamic model according to the basic data of the elevator system; performing dynamic characteristic analysis at each time scale according to the initial theoretical dynamic model and the multi-scale decomposition data set, thereby obtaining multi-scale dynamic characteristic data, wherein the dynamic characteristic analysis at each time scale includes the identification of characteristic parameters at different operation stages, the characteristic analysis of the change of the wire rope stiffness with position, and the evaluation of the change law of the friction force with speed and temperature;

[0053] Step S15: performing a multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator, thereby obtaining multi-scale stability data.

[0054] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in the embodiment of the present invention, step S1 includes the following steps:

[0055] Step S11: acquiring basic data of the elevator system, including structural parameters based on car mass, counterweight mass and wire rope parameters, motor and transmission system parameters, geometry and material parameters of guide rails and guide shoes, and elevator shaft environment parameters;

[0056] The embodiment of the present invention uses sensors and measuring equipment to collect basic data from the elevator system. These data include the mass of the car, the mass of the counterweight, the diameter and material parameters of the wire rope, the power and torque characteristics of the motor, the reduction ratio of the transmission system, the geometric dimensions and material properties of the guide rails, the friction coefficient of the guide shoes, and the temperature, humidity and wind speed in the hoistway environment. Data collection tools include mass sensors, laser rangefinders, strain gauges, and environmental sensors. The basic data obtained undergoes data preprocessing, such as denoising, filtering, and normalization, to ensure the accuracy and consistency of the data. Ultimately, these data will serve as input parameters for subsequent dynamic modeling and analysis.

[0057] Step S12: collecting the position, speed, acceleration, electrical parameters, vibration level and load changes during the operation of the elevator, so as to obtain the historical operation data of the elevator;

[0058] The embodiment of the present invention collects historical data of the elevator during operation through various sensors (such as position sensors, speed sensors, acceleration sensors, current sensors and vibration sensors) installed in the elevator system. These data record the entire process of the elevator from starting, accelerating, running at a constant speed to decelerating and stopping, covering information such as position, speed, acceleration, electrical parameters, vibration level and load change. In order to ensure the integrity of the data, the acquisition system uses a high-speed data acquisition card and storage device. Subsequently, these historical data are pre-processed using signal processing technology, such as low-pass filtering to remove noise, Fourier transform analysis of the spectral characteristics of the vibration signal, and normalization of the data for subsequent analysis.

[0059] Step S13: performing empirical mode decomposition on the elevator historical operation data and decomposing it into components of different time scales, thereby obtaining a multi-scale decomposition data set, wherein the different time scales include high-frequency vibration, medium-frequency motion, and low-frequency trend;

[0060] The historical operation data of the elevator obtained by the embodiment of the present invention uses the empirical mode decomposition (EMD) method to perform multi-scale decomposition on the data. The EMD method decomposes the elevator operation data into different time scale components, mainly including high-frequency vibration components, medium-frequency motion components and low-frequency trend components. In the specific operation, the original data is first normalized, and then the various components from high frequency to low frequency are obtained by gradually extracting the intrinsic mode function (IMF). In each step of the decomposition process, the Hilbert-Huang transform is used to perform instantaneous frequency and energy spectrum analysis on each IMF component to ensure that the physical meaning of the component is clear. Finally, all the decomposed data are combined into a multi-scale decomposition data set for subsequent dynamic characteristics analysis.

[0061] Step S14: establishing an initial theoretical dynamic model according to the basic data of the elevator system; performing dynamic characteristic analysis at each time scale according to the initial theoretical dynamic model and the multi-scale decomposition data set, thereby obtaining multi-scale dynamic characteristic data, wherein the dynamic characteristic analysis at each time scale includes the identification of characteristic parameters at different operation stages, the characteristic analysis of the change of the wire rope stiffness with position, and the evaluation of the change law of the friction force with speed and temperature;

[0062] After completing the generation of the multi-scale decomposition data set, the embodiment of the present invention establishes an initial theoretical dynamic model based on the basic data of the elevator system. The establishment of the model is based on the classical multi-body dynamics theory and the finite element analysis (FEA) method, taking into account the interaction between the rigid body and the flexible body in the elevator system, and including detailed models of the car, counterweight, wire rope, guide rail and transmission system. Static and dynamic simulation analysis of the model is performed using finite element software (such as ANSYS or Abaqus) to obtain the dynamic response of the elevator system at different operating stages. Subsequently, the multi-scale decomposition data set is input into the model, and the dynamic characteristics at each time scale are verified by simulation. The dynamic characteristics analysis includes the identification of characteristic parameters (such as natural frequency, modal vibration shape) at different operating stages, the analysis of the change of wire rope stiffness with the position of the car, and the evaluation of the law of friction with speed and temperature. Finally, all analysis results are integrated to obtain complete multi-scale dynamic characteristics data.

[0063] Step S15: performing a multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator, thereby obtaining multi-scale stability data.

[0064] After completing the multi-scale dynamic characteristic analysis, the embodiment of the present invention uses these characteristic data to perform multi-scale stability analysis. First, based on the multi-scale dynamic characteristic data, stability analysis models under three time scales of high-frequency vibration, medium-frequency motion and low-frequency trend are constructed. These models are parameter identified and calibrated by Lyapunov stability theory and eigenvalue analysis method. In the specific operation, firstly, according to the position, speed and acceleration information in the historical operation data of the elevator, the parameters of the stability analysis model are identified and calibrated by the least squares method or Bayesian estimation. Subsequently, by calculating the eigenvalues, Lyapunov exponents and other indicators of the system at different time scales, the stability boundary of the system is evaluated, and the key parameters and operating conditions that may cause system instability are identified. Finally, using time domain and frequency domain analysis methods, as well as statistical analysis tools (such as correlation analysis, regression analysis, etc.), detailed stability analysis of these key parameters and operating conditions is performed to generate multi-scale stability data; these data will serve as the basis for subsequent coupling sensitivity analysis.

[0065] Preferably, step S15 comprises the following steps:

[0066] Step S151: constructing stability analysis models at different time scales based on the multi-scale dynamic characteristic data, wherein the stability analysis model includes a high-frequency vibration stability model, a medium-frequency motion stability model, and a low-frequency trend stability model;

[0067] After obtaining the multi-scale dynamic characteristic data, the embodiment of the present invention first needs to construct corresponding stability analysis models according to different time scales. Use finite element analysis software (such as ANSYS) and dynamic simulation tools (such as Simulink) to establish stability analysis models for high-frequency vibration, medium-frequency motion and low-frequency trend respectively. In the process of model construction, appropriate boundary conditions and initial conditions are selected to ensure that the model can accurately reflect the actual dynamic behavior of the elevator system. For the high-frequency vibration model, special attention is paid to the friction between the guide rail and the guide shoe and the elastic restoring force of the wire rope; the medium-frequency motion model mainly analyzes the mass distribution of the car and the counterweight and its influence on the motor output; the low-frequency trend model focuses on the structural deformation and the influence of ambient temperature on the stability of the system during the long-term operation of the elevator. Through these models, an accurate simulation platform can be provided for subsequent stability analysis.

[0068] Step S152: Parameter identification and calibration of the stability analysis model are performed according to the position, speed and acceleration information in the historical operation data of the elevator, and the stability boundary of the system is evaluated based on different time scales to obtain stability boundary data, wherein the evaluation based on different time scales includes calculating the characteristic value and Lyapunov exponent of the stability boundary of the system at each time scale;

[0069] After the stability analysis model is constructed, the embodiment of the present invention uses the historical operation data of the elevator to identify and calibrate the parameters of the model; first, according to the model characteristics of different time scales, key parameters such as position, velocity and acceleration are selected. These parameters are preliminarily identified using the least squares method and Kalman filter, and then the parameters are further calibrated through global optimization algorithms such as genetic algorithm (GA) or particle swarm optimization (PSO) to improve the accuracy of the model. Through iterative calculation, the parameters of the model are optimized to minimize the error between the simulation results and the historical data. Finally, a calibrated stability analysis model is obtained, and the stability boundary data of the system at different time scales is calculated to provide a basis for subsequent stability analysis.

[0070] Step S153: identifying key parameters and operating conditions that may cause system instability in the stability analysis model, thereby obtaining key parameter data;

[0071] The embodiment of the present invention uses a calibrated stability analysis model to identify key parameters that may cause system instability based on the position information, speed information and acceleration information in the historical operation data of the elevator. First, a sensitivity analysis method is used to analyze the influence of each parameter on the system stability at different time scales. Then, the system stability boundary at each time scale is determined by Lyapunov exponent and eigenvalue calculation. In this process, computing tools such as MATLAB are used to conduct a detailed evaluation of the stability of the system to identify critical parameter values ​​that may cause instability under different operating conditions. Finally, these key parameter data are used as input, combined with stability boundary data, to provide a basis for subsequent time domain and frequency domain stability analysis.

[0072] Step S154: Perform time domain and frequency domain stability analysis based on key parameter data and stability boundary data, and use statistical methods to perform stability analysis on the elevator historical operation data to generate multi-scale stability data.

[0073] The embodiment of the present invention uses time domain and frequency domain analysis methods to conduct an in-depth analysis of the stability of the system based on key parameter data and system stability boundary data. First, based on time domain analysis, the transient response of the system at each time scale is calculated using the historical operation data of the elevator, with special attention paid to the response characteristics when the system is close to the stability boundary. The data is analyzed in the frequency domain using the short-time Fourier transform (STFT) to detect the potential unstable frequency components in the system. By combining the time domain and frequency domain results, the key factors affecting the stability of the system are identified, and the degree of influence of these factors on the stability of the system is verified using statistical methods (such as regression analysis and correlation analysis). Finally, detailed multi-scale stability data is generated to lay the foundation for subsequent subsystem coupling sensitivity analysis. After completing the time domain and frequency domain stability analysis, the multi-scale stability data is generated by combining the stability analysis models at different time scales. Specifically, the results of the time domain and frequency domain analysis are first fused to extract the key stability indicators at each time scale. Multi-scale data are integrated using a data fusion algorithm (such as Bayesian fusion or weighted average), and the data is normalized to ensure that the data at different time scales are consistent and comparable. Finally, the multi-scale stability data are organized into a complete data set, including stability boundaries, key parameters and their influencing factors at high, medium and low frequencies, providing necessary data support for subsequent subsystem coupling sensitivity analysis and global nonlinear characteristic mapping processing.

[0074] Preferably, step S2 comprises the following steps:

[0075] Step S21: establishing a multi-subsystem model of the elevator system according to the basic data of the elevator system, wherein the multi-subsystem model includes a car subsystem, a counterweight subsystem, a wire rope subsystem, a drive subsystem and a guide subsystem;

[0076] Step S22: constructing a coupling relationship matrix between subsystems reflecting the degree of mutual influence between different subsystems based on the multi-scale stability data and the multi-subsystem model;

[0077] Step S23: using a global sensitivity analysis method to identify the inter-subsystem coupling factors that have the greatest impact on system stability in the coupling relationship matrix, thereby obtaining stability impact data;

[0078] Step S24: performing parameter change tests on the stability influencing data based on the orthogonal experimental design method, analyzing the influence of the parameter changes on the system stability, and establishing a mapping relationship model between the subsystem coupling parameters and the system stability using support vector regression;

[0079] Step S25: calculating the sensitivity coefficient of each subsystem coupling parameter to the system stability in the mapping relationship model, and ranking the subsystem coupling parameters in importance according to the sensitivity coefficients, thereby obtaining a list of key coupling parameters;

[0080] Step S26: construct a multiscale coupling sensitivity model that reflects the impact of subsystem coupling on system stability at different time scales based on the key coupling parameter list, and perform uncertainty analysis and optimization on the multiscale coupling sensitivity model to generate multiscale coupling sensitivity data.

[0081] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment of the present invention, step S2 includes the following steps:

[0082] Step S21: establishing a multi-subsystem model of the elevator system according to the basic data of the elevator system, wherein the multi-subsystem model includes a car subsystem, a counterweight subsystem, a wire rope subsystem, a drive subsystem and a guide subsystem;

[0083] The embodiment of the present invention establishes a multi-subsystem model of the elevator based on the basic data of the elevator system using simulation software (such as Simulink or ANSYS). By analyzing the overall structure of the elevator, the system is divided into multiple subsystems such as a car subsystem, a counterweight subsystem, a wire rope subsystem, a drive subsystem, and a guide subsystem. For the car subsystem, its mass distribution, stiffness, and inertia parameters are considered in particular; the counterweight subsystem needs to analyze its mass and equilibrium state; the wire rope subsystem involves parameters such as the elastic modulus and friction characteristics of the wire rope; the modeling of the drive subsystem is based on the characteristic curve of the motor and the efficiency of the transmission device; the guide subsystem mainly considers the contact characteristics and friction between the guide rail and the guide shoe. Through these steps, the dynamic models of each subsystem are constructed and integrated into a unified elevator system simulation model, laying the foundation for subsequent subsystem coupling analysis.

[0084] Step S22: constructing a coupling relationship matrix between subsystems reflecting the degree of mutual influence between different subsystems based on the multi-scale stability data and the multi-subsystem model;

[0085] After completing the establishment of the multi-subsystem model, the embodiment of the present invention constructs a coupling relationship matrix between subsystems based on multi-scale stability data using matrix analysis tools (such as the matrix operation function in MATLAB). First, each element of the coupling relationship matrix is ​​defined according to the mutual influence between the output and input of each subsystem model. By simulating the models of each subsystem, the coupling effects between different subsystems are measured and quantified. For example, the influence of the wire rope subsystem on the car subsystem and the interaction between the drive subsystem and the counterweight subsystem are analyzed. The generated coupling relationship matrix reflects the degree of mutual coupling between the subsystems and is used to identify the coupling factors that have the greatest impact on the stability of the system.

[0086] Step S23: using a global sensitivity analysis method to identify the inter-subsystem coupling factors that have the greatest impact on system stability in the coupling relationship matrix, thereby obtaining stability impact data;

[0087] The embodiment of the present invention uses a global sensitivity analysis method to identify the coupling factors of each subsystem in the coupling relationship matrix and determine the coupling factors that have the greatest impact on the stability of the system. By designing a sensitivity analysis experiment, the coupling parameters of different subsystems are disturbed, and the impact of these disturbances on the overall stability of the system is observed. The sensitivity coefficients of each coupling parameter are calculated using Monte Carlo simulation or Sobol sensitivity analysis. These sensitivity coefficients can quantify the importance of each coupling factor to the stability of the system and rank them, thereby identifying the most critical coupling factors in the system, generating stability impact data, and laying the foundation for subsequent parameter change experiments.

[0088] Step S24: performing parameter change tests on the stability influencing data based on the orthogonal experimental design method, analyzing the influence of the parameter changes on the system stability, and establishing a mapping relationship model between the subsystem coupling parameters and the system stability using support vector regression;

[0089] After the embodiment of the present invention identifies the key coupling factors, an orthogonal experimental design method is used to perform parameter change experiments to deeply analyze the impact of these factors on system stability. An experimental scheme is designed to analyze the impact of different combinations of key coupling parameters on the dynamic behavior of the system by performing system simulation. A system simulation tool (such as Simulink) is used to simulate the operation of the elevator under different parameter combinations, and the stability data of each experiment is recorded. Then, the support vector regression (SVR) algorithm is used to perform nonlinear fitting on the experimental data to establish a mapping relationship model between key coupling parameters and system stability. This model can be used to predict the stability of the system under different parameter settings, providing a basis for subsequent sensitivity models.

[0090] Step S25: calculating the sensitivity coefficient of each subsystem coupling parameter to the system stability in the mapping relationship model, and ranking the subsystem coupling parameters in importance according to the sensitivity coefficients, thereby obtaining a list of key coupling parameters;

[0091] The embodiment of the present invention calculates the sensitivity coefficient of the coupling parameters of each subsystem based on the established mapping relationship model. Statistical analysis software (such as Scikit-learn in MATLAB or Python) is used to perform sensitivity analysis on each coupling parameter to quantify its impact on system stability. The importance of the parameters is ranked by the calculated sensitivity coefficients to determine the coupling parameters that are most critical to system stability. This process ensures that these highly sensitive parameters can be focused on during subsequent optimization, thereby improving the overall stability and reliability of the system.

[0092] Step S26: construct a multiscale coupling sensitivity model that reflects the impact of subsystem coupling on system stability at different time scales based on the key coupling parameter list, and perform uncertainty analysis and optimization on the multiscale coupling sensitivity model to generate multiscale coupling sensitivity data.

[0093] According to the aforementioned sensitivity analysis results, the embodiment of the present invention constructs a multi-scale coupling sensitivity model that reflects the influence of subsystem coupling on system stability at different time scales. First, the key coupling parameter data is combined with the multi-scale stability data, and the coupling sensitivity model at each time scale is constructed using multivariate regression analysis technology. Then, uncertainty analysis methods (such as Monte Carlo simulation or Latin hypercube sampling) are used to test the robustness of the model, evaluate the performance of the model under parameter changes, and optimize the model structure to reduce uncertainty. Finally, the generated multi-scale coupling sensitivity data provides accurate parameter support for further global nonlinear dynamic model correction.

[0094] Preferably, step S3 comprises the following steps:

[0095] Step S31: Analyze the multi-scale coupling sensitivity data using a clustering algorithm to identify and classify the main nonlinear characteristics in the system, thereby obtaining nonlinear characteristic classification data;

[0096] The embodiment of the present invention first inputs the multi-scale coupling sensitivity data into the clustering algorithm for processing; commonly used clustering algorithms such as K-means, DBSCAN or hierarchical clustering can be used for this purpose. Through these algorithms, various types of nonlinear characteristics in the elevator system are identified and classified, such as the nonlinear elastic characteristics of the wire rope, the nonlinear characteristics of friction, and the nonlinear coupling relationship between different subsystems. The nonlinear characteristic classification data formed by the clustering results is used for further nonlinear function fitting and selection. The implementation of this step ensures that the nonlinear behavior of the system can be accurately reflected in the subsequent model.

[0097] Step S32: performing nonlinear function fitting and selection based on the nonlinear characteristics of the nonlinear characteristic classification data, thereby constructing a nonlinear mapping function set;

[0098] The embodiment of the present invention uses nonlinear regression or machine learning algorithms (such as support vector regression SVR, neural network or Gaussian process regression) to fit and select nonlinear functions based on the nonlinear characteristic classification data generated in the previous step. By taking the classified data as input, the corresponding nonlinear function is fitted to construct a nonlinear mapping function set. This function set can accurately describe various types of nonlinear behaviors in the elevator system, such as the change of friction with speed and temperature, or the change of wire rope stiffness with position. Ultimately, the constructed nonlinear mapping function set will be used to integrate into the global nonlinear dynamic model to provide an accurate nonlinear dynamic description for the entire system.

[0099] Step S33: integrating the nonlinear mapping function set according to the preset basic dynamic equation of the elevator system, thereby constructing a global nonlinear dynamic model;

[0100] After the nonlinear mapping function set is constructed, the embodiment of the present invention integrates these nonlinear functions into the global model according to the basic dynamic equations of the elevator system to construct a global nonlinear dynamic model. By coupling the nonlinear behavior of each subsystem to the overall dynamic equation, a numerical solution method (such as finite element method FEM or multi-body dynamics MBD simulation) can be used to solve this nonlinear dynamic model. This global model can accurately reflect the dynamic characteristics of the elevator system under different operating conditions, and provides a basis for subsequent parameter identification.

[0101] Step S34: designing a parameter identification model based on a machine learning algorithm according to the global nonlinear dynamics model, and designing an adaptive learning rate strategy to optimize the convergence performance of the parameter identification model, thereby generating an adaptive parameter identification strategy;

[0102] According to the global nonlinear dynamic model, the embodiment of the present invention first performs parameter sensitivity analysis to identify the key parameters that have the greatest impact on the model behavior. This step can be performed using statistical analysis tools (such as sensitivity analysis libraries in MATLAB or Python). Then, using these sensitivity data, a parameter identification model based on a machine learning algorithm, such as linear regression, random forest or neural network, is designed. In order to optimize the convergence performance of parameter identification, an adaptive learning rate strategy is designed to ensure that the model can quickly and accurately converge to the optimal parameter value during the training process. Finally, an adaptive parameter identification strategy is generated to ensure that the model can adaptively adjust parameters in practical applications to adapt to different operating conditions.

[0103] Step S35: performing time series analysis on the multi-scale stability data based on the historical operation data of the elevator, and using the sliding time window method to identify the time-varying characteristics of the system parameters, thereby obtaining the parameter time-varying characteristic data;

[0104] The embodiment of the present invention is based on the historical operation data of the elevator. First, time series analysis (such as ARIMA model, LSTM neural network or wavelet analysis) is performed to extract the time-varying characteristic data of the system parameters. Then, the sliding time window method is used to apply these time-varying characteristic data to system parameter identification to identify the trend and law of parameter changes over time. Next, a time-varying parameter estimation algorithm is designed based on the Kalman filter algorithm, and a time-varying parameter update strategy is generated by continuously updating the state estimation of the system. This strategy includes a parameter framework update strategy, a parameter update frequency strategy, a parameter smooth transition strategy, and a parameter update constraint strategy to ensure that the system parameters can be updated in a timely and accurate manner at different operation stages.

[0105] Step S36: Design a time-varying parameter estimation algorithm based on Kalman filtering according to the parameter time-varying characteristic data, and use the time-varying parameter estimation algorithm to generate an update strategy for the parameter time-varying characteristic data, thereby obtaining a time-varying parameter update strategy, wherein the update strategy specifically includes a parameter framework update strategy, a parameter update frequency strategy, a parameter smooth transition strategy, and a parameter update constraint strategy.

[0106] After determining the time-varying parameter update strategy, the embodiment of the present invention designs a time-varying parameter estimation algorithm based on Kalman filtering, which specifically includes the linearization of the state equation and the observation equation, and the setting of the process noise and the observation noise covariance matrix. Then, in combination with the recursive least squares (RLS) algorithm, real-time estimation of the parameters is performed to generate time-varying estimation data of the parameters. Based on these data, algorithm switching strategy data is generated for switching between different algorithms (such as RLS and EKF) to optimize the accuracy and efficiency of the estimation. Finally, the rate of change of the parameters is constrained by the penalty function method to generate constraint processing mechanism data. All generated strategies are integrated to design a parameter identification model based on machine learning, and the convergence performance of the model is further optimized by the adaptive learning rate strategy to generate the final adaptive parameter identification strategy.

[0107] Preferably, step S34 includes the following steps:

[0108] Step S341: performing parameter sensitivity analysis to determine the key parameters that have the greatest impact on the model behavior based on the global nonlinear dynamics model, thereby obtaining parameter sensitivity data;

[0109] The embodiment of the present invention uses a global nonlinear dynamic model and a sensitivity analysis tool (such as Simulink in MATLAB or SALib library in Python) to perform sensitivity analysis on the key parameters of the model. Specifically, by changing the value of each parameter in the model and observing its impact on the system behavior, the sensitivity index of each parameter is calculated. The generated parameter sensitivity data is used to determine which parameters have the greatest impact on the dynamic behavior of the model. These key parameters will be identified and optimized first in subsequent steps to ensure that the model can accurately reflect the dynamic characteristics of the elevator system.

[0110] Step S342: transforming the global nonlinear dynamics model into a mathematical expression for parameter identification based on key parameters according to the parameter sensitivity data, thereby obtaining a parameter identification problem expression;

[0111] The embodiment of the present invention converts the global nonlinear dynamic model into a parameter identification mathematical expression based on key parameters based on the parameter sensitivity data generated in the previous step. Specifically, the parameters with high sensitivity are incorporated into the identification equation, and the objective function and constraint conditions of parameter identification are established. This expression can be solved by using the least squares method or other optimization algorithms to finally obtain the mathematical expression of the parameter identification problem. This step ensures that the most important parameters in the model can be accurately identified and estimated.

[0112] Step S343: setting initial values ​​for the parameters to be identified according to the parameter identification problem expression and the preset prior knowledge, and performing recursive least squares algorithm design and extended Kalman filter algorithm design, so as to obtain RLS and EKF algorithm data, wherein the recursive least squares algorithm design specifically includes forgetting factor determination and covariance matrix initialization, wherein the extended Kalman filter algorithm design specifically includes linearization of state equations and observation equations and setting of process noise and observation noise covariance matrices;

[0113] The embodiment of the present invention first sets initial values ​​for the parameters to be identified, and these initial values ​​can be determined based on preset prior knowledge or historical operation data. Then, a recursive least squares (RLS) algorithm is designed, which includes the determination of the forgetting factor and the initialization of the covariance matrix. This part of the design can be implemented through the relevant libraries of MATLAB or Python. At the same time, an extended Kalman filter (EKF) algorithm is designed, including the linearization of the state equation and the observation equation and the setting of the process noise and observation noise covariance matrix. Finally, RLS and EKF algorithm data are obtained, which will be used in parameter estimation and model optimization in subsequent steps.

[0114] Step S344: generating an algorithm switching strategy based on the estimation error according to the RLS and EKF algorithm data, thereby obtaining the algorithm switching strategy data;

[0115] The embodiment of the present invention designs an algorithm switching strategy based on the estimation error based on the previously generated RLS and EKF algorithm data. By comparing the estimation errors of the two algorithms under different conditions, it is decided when to switch from one algorithm to another to ensure the optimal estimation accuracy and computational efficiency. This strategy can be implemented by writing adaptive control logic and debugging and verifying it in combination with actual operation data. The generated algorithm switching strategy data will be dynamically applied in the process of parameter identification to improve the accuracy and robustness of the overall model.

[0116] Step S345: setting a value range and limiting a change rate for the parameter to be identified according to the algorithm switching strategy and the parameter identification problem expression, and designing a parameter constraint processing mechanism by a penalty function method, thereby obtaining constraint processing mechanism data;

[0117] The embodiment of the present invention sets the value range and limits the change rate of the parameters to be identified based on the algorithm switching strategy and the parameter identification problem expression. The penalty function method is used to design a processing mechanism for the parameter constraint conditions to ensure that the parameters will not exceed the physically feasible range or undergo abnormal changes during the identification process. This step ensures the stability and reliability of the parameter estimation process by optimizing the constraint condition design in the algorithm. The generated constraint processing mechanism data will be applied to the parameter identification model to ensure that the model can maintain good estimation performance under different working conditions.

[0118] Step S346: Design a parameter identification model based on the machine learning algorithm according to the algorithm switching strategy data, constraint processing mechanism data, and RLS and EKF algorithm data, and design an adaptive learning rate strategy to optimize the convergence performance of the parameter identification model, thereby generating an adaptive parameter identification strategy.

[0119] The embodiment of the present invention combines the algorithm switching strategy data, constraint processing mechanism data, and RLS and EKF algorithm data generated above to design a parameter identification model based on a machine learning algorithm; this model can be implemented by a deep learning algorithm (such as LSTM, GRU, or an adaptive algorithm based on a neural network), and by designing an adaptive learning rate strategy, it is ensured that the model can converge quickly under different operating conditions. Finally, through continuous iteration and optimization, an adaptive parameter identification strategy is generated. This strategy will be used to update the dynamic model of the elevator system online to ensure that the model always maintains a high-precision prediction of the actual system behavior.

[0120] Preferably, step S4 comprises the following steps:

[0121] Step S41: acquiring real-time status monitoring data of the elevator system by deploying a sensor network, and performing preprocessing on the real-time status monitoring data based on denoising, filtering and data standardization, thereby obtaining real-time status data;

[0122] The embodiment of the present invention collects the state data of the elevator system in real time by deploying a sensor network (such as accelerometers, velocity sensors, temperature sensors, etc.) at key parts of the elevator system. These sensor data need to be pre-processed by denoising, filtering and data standardization. Specifically, a low-pass filter, a Kalman filter or a wavelet denoising technique can be used to remove sensor noise. Then, the data is standardized to make it suitable for subsequent machine learning and anomaly detection algorithms. Data preprocessing is performed through tools such as MATLAB and Python to ensure that the acquired data is accurate and reliable, and to lay the foundation for the identification of abnormal states.

[0123] Step S42: using a machine learning method to establish a normal operation mode of the elevator system and a feature library of various abnormal states according to historical operation data, thereby obtaining an abnormal state feature library;

[0124] The embodiment of the present invention uses machine learning methods (such as support vector machine SVM, decision tree, deep learning algorithm, etc.) based on historical operation data to establish the normal operation mode of the elevator system and the feature library of various abnormal states. This step includes clustering analysis, anomaly detection and classification modeling of historical data. Through feature extraction technology (such as principal component analysis PCA, independent component analysis ICA), representative features are extracted from the data to construct a feature library of normal and abnormal states. The feature library is used for subsequent abnormal state detection and provides basic data for classification.

[0125] Step S43: comparing the real-time status data with the abnormal status feature library, and identifying potential abnormal status, thereby generating abnormal warning signal data;

[0126] The embodiment of the present invention compares the real-time status data with the abnormal state feature library and uses a classification algorithm (such as the k-nearest neighbor algorithm KNN, support vector machine SVM or deep learning model) to identify potential abnormal states. First, the real-time data is converted into a feature vector and matched with the abnormal state pattern in the feature library. Then, by calculating the distance metric (such as Euclidean distance or Mahalanobis distance), it is determined whether the real-time data meets the characteristics of the abnormal state. Abnormal warning signal data is generated based on the classification results to prompt the system of possible abnormal conditions.

[0127] Step S44: activating corresponding modules according to abnormal warning signal data and adaptive parameter identification strategy, and quickly estimating short-term dynamic changes of the system, thereby generating short-term dynamic correction parameters;

[0128] According to the identified abnormal warning signal data, the embodiment of the present invention activates the corresponding system module to quickly estimate the short-term dynamic changes. Time series prediction algorithms (such as ARIMA, LSTM neural network) can be used to perform short-term dynamic predictions to generate short-term dynamic correction parameters. These parameters are used to adjust the real-time control strategy of the system to cope with dynamic changes in the short term. Through model prediction and parameter adjustment, it is ensured that the elevator system can respond quickly and correct abnormal conditions to maintain system stability.

[0129] Step S45: evaluating the long-term trend of the system based on the time-varying parameter update strategy, the abnormal warning signal data, and the multi-scale dynamic characteristic data, thereby generating a long-term trend correction parameter;

[0130] The embodiment of the present invention evaluates the long-term change trend of the system based on the time-varying parameter update strategy, abnormal warning signal data and multi-scale dynamic characteristic data. This step can be achieved through long time series analysis (such as sliding average method, trend analysis) and multi-dimensional regression analysis. By analyzing the trend of long-term data, long-term trend correction parameters are generated. These parameters are used to adjust the long-term control strategy of the system to cope with changes and potential abnormalities of the elevator system in long-term operation.

[0131] Step S46: Perform weighted fusion of short-term dynamic correction parameters and long-term trend correction parameters based on system changes at multiple time scales to generate comprehensive dynamic correction parameter data.

[0132] The embodiment of the present invention performs weighted fusion of short-term dynamic correction parameters and long-term trend correction parameters to generate comprehensive dynamic correction parameters. This step can be implemented using weighted averaging, fuzzy logic control or weighted fusion algorithm to comprehensively consider short-term and long-term correction parameters to form the final dynamic correction parameters. The comprehensive correction parameters will be used for real-time adjustment and optimization of the system to ensure that the elevator system can maintain good performance and stability at multiple time scales.

[0133] Preferably, step S5 comprises the following steps:

[0134] Step S51: applying the dynamic correction parameter data to the global nonlinear dynamic model, and updating the model parameters and the model structure, thereby generating an updated dynamic model;

[0135] The embodiment of the present invention applies the dynamic correction parameters generated from step S46 to the global nonlinear dynamic model. The specific operation includes integrating the correction parameters into the parameter settings of the model and updating the model parameters and structure through programming. The model library in MATLAB Simulink or Python can be used to load the updated parameters and reconfigure the model of the elevator system. Then, the model simulation is rerun to generate an updated dynamic model that can reflect the latest system characteristics and correction data.

[0136] Step S52: using the updated dynamics model to predict the short-term behavior of the elevator system, thereby generating model prediction result data;

[0137] The embodiment of the present invention uses the updated dynamic model to predict the short-term behavior of the elevator system, and uses model predictive control (MPC) technology or other prediction algorithms (such as ARIMA and LSTM networks in time series analysis) to predict the behavior of the system in the short term. The updated model is applied to the simulation environment to predict future operating conditions and generate model prediction result data. These prediction results will help analyze the performance of the elevator system in the upcoming working cycle and provide basic data for the subsequent evaluation of the model correction effect.

[0138] Step S53: Calculate statistical indicators of the model prediction result data and the real-time status monitoring data, thereby generating model accuracy evaluation indicator data;

[0139] The embodiment of the present invention compares the generated model prediction result data with the actual real-time state monitoring data to calculate statistical indicators; this includes calculating the error between the model prediction result and the actual data (such as mean square error MSE, mean absolute error MAE). By using statistical methods and error analysis tools (such as Python's Scikit-learn library or the statistical toolbox in MATLAB), model accuracy evaluation index data is generated. These indicators will be used to evaluate the prediction accuracy and reliability of the updated dynamic model.

[0140] Step S54: evaluating the effectiveness of correcting the updated dynamics model according to the preset performance threshold and model accuracy evaluation index data, thereby generating model correction effect data;

[0141] The embodiment of the present invention evaluates the correction effect of the updated dynamic model based on the model accuracy evaluation index data and the pre-set performance threshold; this step includes setting a specific performance threshold (such as error tolerance) and comparing it with the model accuracy evaluation index. Decision tree analysis, Bayesian optimization or other optimization algorithms are used to determine whether the correction effect of the model meets expectations; the generated model correction effect data will be used to evaluate the effectiveness of the correction work and ensure the accuracy and stability of the model in practical applications.

[0142] Step S55: adjusting the learning rate and forgetting factor in the adaptive parameter identification strategy based on the model correction effect data, thereby generating a parameter identification optimization strategy;

[0143] The embodiment of the present invention adjusts the learning rate and forgetting factor in the adaptive parameter identification strategy based on the model correction effect data. An optimization algorithm (such as particle swarm optimization PSO, genetic algorithm GA) is used to optimize the learning rate and forgetting factor to improve the accuracy of parameter identification and the convergence speed of the model. The specific operation includes adjusting the hyperparameter settings in the parameter identification model and performing iterative testing to determine the optimal configuration. The parameter identification optimization strategy finally generated will be used to further improve the parameter adjustment process of the model.

[0144] Step S56: structurally optimizing the updated dynamics model according to the parameter identification optimization strategy and the model correction effect data, thereby generating optimized model structure data.

[0145] The embodiment of the present invention optimizes the structure of the updated dynamic model according to the parameter identification optimization strategy and the model correction effect data. This includes redesigning the model structure, increasing or decreasing the parameters in the model, improving the mathematical expression of the model, etc. The model structure is adjusted using a model structure optimization algorithm (such as a structure optimization algorithm, a model selection technique), and optimized model structure data is generated. The optimized model structure data finally obtained will be used for further system design and debugging to ensure that the dynamic model of the elevator system performs best in practical applications.

[0146] Preferably, step S6 comprises the following steps:

[0147] Step S61: using the empirical mode decomposition method to perform high-frequency vibration mode recognition in the elevator system on the multi-scale dynamic characteristic data, thereby obtaining high-frequency vibration mode recognition data;

[0148] The embodiment of the present invention uses the empirical mode decomposition (EMD) method to perform high-frequency vibration pattern recognition on the multi-scale dynamic characteristic data of the elevator system. The specific operation includes using Python's PyEMD library or MATLAB's EMD toolbox to decompose the signal into different intrinsic mode functions (IMFs). High-frequency vibration components are extracted from these IMFs, feature analysis is performed, and patterns related to high-frequency vibration are identified. These pattern recognition data will be used for subsequent state observer design.

[0149] Step S62: integrating the high-frequency vibration pattern recognition data into the state space model based on the optimized model structure data, and constructing an extended state space model;

[0150] The embodiment of the present invention integrates the high-frequency vibration pattern recognition data into the state space model based on the optimization model structure data to construct an extended state space model. This process includes using the identified high-frequency vibration pattern as a state variable, combining it with the optimization model structure data, and constructing a new state space equation. Specifically, the state space model can be constructed using MATLAB's State-Space toolbox or Python's control library. The generated extended state space model will be used to design an extended state observer.

[0151] Step S63: performing an extended state observer design according to the extended state space model, thereby obtaining an extended state observer design scheme, wherein the extended state observer design includes determining an observer gain matrix and processing nonlinear terms and uncertainties;

[0152] The embodiment of the present invention designs an extended state observer (ESO) according to an extended state space model; first, the observer gain matrix is ​​determined, which can be achieved by an optimization algorithm (such as LQR or H∞ control design) to ensure the observer's estimation accuracy of the system state. Secondly, to deal with nonlinear terms and system uncertainties, nonlinear observer design techniques (such as sliding mode observers or adaptive observers) can be used. The specific design steps can be implemented using MATLAB's Control System Toolbox or Python's scipy library. Finally, an extended state observer design scheme is obtained for accurate estimation of the state.

[0153] Step S64: using an extended state observer design scheme to estimate the state that is difficult to measure directly, thereby obtaining high-frequency vibration estimation data;

[0154] The embodiment of the present invention uses an extended state observer design to estimate the state in the elevator system that is difficult to measure directly. This includes applying the extended state observer to real-time data to estimate state variables that are difficult to measure directly, such as high-frequency vibration. Using the designed extended state observer, the measurement data and the estimation results are fused through the observer's feedback mechanism to generate high-frequency vibration estimation data. These estimation data will be used to further optimize the collaborative optimization of the subsystem.

[0155] Step S65: performing multi-subsystem collaborative optimization according to the high-frequency vibration estimation data and the multi-scale coupling sensitivity data to generate coupling optimization parameters between subsystems;

[0156] The embodiment of the present invention performs collaborative optimization of multiple subsystems based on high-frequency vibration estimation data and multi-scale coupling sensitivity data. This can be achieved through optimization algorithms (such as multi-objective optimization, co-evolutionary algorithms) to ensure that the coupling relationship between different subsystems is optimized. Specifically, the estimated data and sensitivity information are input into the optimization model, and collaborative adjustments are made to improve the performance of the overall system. The optimization process can be performed using the optimization toolbox in MATLAB or Python to generate coupling optimization parameters between subsystems.

[0157] Step S66: Correcting the global nonlinear dynamics model according to the inter-subsystem coupling optimization parameters to obtain a corrected elevator dynamics model.

[0158] The embodiment of the present invention corrects the global nonlinear dynamic model according to the coupling optimization parameters between subsystems. This includes applying the optimized subsystem coupling parameters to the global model and readjusting the nonlinear characteristics of the model. The specific operation includes modifying the nonlinear terms of the model, updating the parameter settings of the model, and verifying the effect of the model correction through simulation. The model is corrected using MATLAB Simulink or the scipy.optimize tool in Python to obtain a corrected elevator dynamic model.

[0159] The present invention also provides an elevator dynamics model correction system, which is used to execute the above-mentioned elevator dynamics model correction method, and the elevator dynamics model correction system comprises:

[0160] The basic data and multi-scale analysis module is used to obtain the basic data of the elevator system and the historical operation data of the elevator; perform multi-scale dynamic characteristics analysis based on the basic data of the elevator system to generate multi-scale dynamic characteristics data; perform multi-scale stability analysis on the multi-scale dynamic characteristics data based on the historical operation data of the elevator to obtain multi-scale stability data;

[0161] A coupling sensitivity analysis module is used to perform subsystem coupling sensitivity analysis on multi-scale stability data according to structural parameters in the basic data of the elevator system, thereby generating multi-scale coupling sensitivity data;

[0162] The nonlinear model and parameter identification module is used to perform global nonlinear characteristic mapping processing based on multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; perform parameter identification strategy processing based on the global nonlinear dynamic model to generate an adaptive parameter identification strategy; perform time-varying characteristic identification processing on multi-scale stability data based on the historical operation data of the elevator to generate a time-varying parameter update strategy;

[0163] The real-time monitoring and dynamic correction module is used to obtain real-time status monitoring data of the elevator system, identify abnormal status based on the real-time status monitoring data, and obtain abnormal warning signal data; use adaptive parameter identification strategy and time-varying parameter update strategy to perform multi-time scale dynamic correction processing on abnormal warning signal data to obtain dynamic correction parameter data;

[0164] The model updating and optimization module is used to perform model updating according to the dynamic correction parameter data, evaluate the model accuracy, and generate model correction effect data; based on the model correction effect data, the model structure is re-optimized according to the adaptive parameter identification strategy to generate optimized model structure data;

[0165] The extended observation and correction module is used to construct an extended state observer based on the high-frequency vibration mode according to the optimization model structure data, and use the extended state observer to estimate the state that is difficult to measure directly, so as to obtain high-frequency vibration estimation data; perform multi-subsystem collaborative optimization based on the high-frequency vibration estimation data and multi-scale coupling sensitivity data to generate inter-subsystem coupling optimization parameters; and correct the global nonlinear dynamic model according to the inter-subsystem coupling optimization parameters to obtain the corrected elevator dynamic model.

[0166] The present invention integrates basic data and historical operation data to conduct multi-scale dynamics and stability analysis; it provides a basis for in-depth understanding of the operating status of the elevator system, can identify the behavioral characteristics and stability status of the system at different time scales, and lay a solid foundation for subsequent analysis and optimization. Based on structural parameters and stability data, the coupling sensitivity between subsystems is analyzed; the influence of the interaction between different subsystems on the overall stability of the system is revealed, and the key factors affecting the performance of the system are identified, which provides a basis for the fine adjustment and optimization of the system. It processes global nonlinear characteristics, constructs dynamic models, and develops parameter identification and time-varying characteristic identification strategies; by constructing accurate nonlinear dynamic models and adaptive parameter identification strategies, the model's ability to predict system behavior is improved, and it can adapt to changes in system operation, enhancing the practicality and flexibility of the model. Real-time monitoring of system status, identification of anomalies, and dynamic correction of multiple time scales are performed; through real-time monitoring and anomaly warning, potential problems can be discovered and responded to in a timely manner, and the dynamic correction strategy ensures that the system can maintain good performance under various working conditions, improving the reliability and safety of the system. Update the model according to the correction parameters, evaluate the model accuracy, and perform structural optimization; the continuous updating and optimization of the model ensures the accuracy and adaptability of the model, making the design and operation of the elevator system more in line with actual needs and improving the efficiency and performance of the system. Build an extended state observer to estimate the difficult-to-measure state, perform multi-subsystem collaborative optimization and dynamic model correction; through high-frequency vibration estimation and coupling optimization between subsystems, the understanding and control capabilities of the complex dynamic behavior of the system are further improved. The corrected model is more accurate and can better guide the design and maintenance of the elevator system. Overall, this process forms a closed-loop system optimization and control process from data collection to model construction, parameter identification, real-time monitoring, dynamic correction, and then to model update and optimization. The output of each module provides input for the next module, ensuring the consistency and effectiveness of the entire system analysis and optimization; through this systematic approach, the performance, reliability and safety of the elevator system can be significantly improved.

[0167] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0168] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for correcting an elevator dynamics model, characterized in that: The following steps are involved: Step S1: acquiring basic data of the elevator system and historical operation data of the elevator; performing multi-scale dynamic characteristic analysis according to the basic data of the elevator system to generate multi-scale dynamic characteristic data; performing multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator to obtain multi-scale stability data; Step S2: performing subsystem coupling sensitivity analysis on the multi-scale stability data according to the structural parameters in the elevator system basic data, thereby generating multi-scale coupling sensitivity data; Step S3: performing global nonlinear characteristic mapping processing according to the multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; Process the parameter identification strategy according to the global nonlinear dynamics model and generate an adaptive parameter identification strategy; Based on the historical operation data of the elevator, the time-varying characteristics of the multi-scale stability data are identified and processed to generate a time-varying parameter update strategy; Step S4: acquiring real-time status monitoring data of the elevator system, identifying abnormal status based on the real-time status monitoring data, and obtaining abnormal warning signal data; performing multi-time scale dynamic correction processing on the abnormal warning signal data using an adaptive parameter identification strategy and a time-varying parameter update strategy, and obtaining dynamic correction parameter data; Step S5: performing model update according to the dynamic correction parameter data, and performing model accuracy evaluation to generate model correction effect data; re-optimizing the model structure according to the adaptive parameter identification strategy based on the model correction effect data to generate optimized model structure data; Step S6: constructing an extended state observer based on the high-frequency vibration mode according to the optimized model structure data, and using the extended state observer to estimate the state that is difficult to measure directly, thereby obtaining high-frequency vibration estimation data; Based on the high-frequency vibration estimation data and multi-scale coupling sensitivity data, multi-subsystem collaborative optimization is performed to generate coupling optimization parameters between subsystems; The global nonlinear dynamic model is corrected according to the coupling optimization parameters between subsystems to obtain a corrected elevator dynamic model.

2. The elevator dynamics model correction method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: acquiring basic data of the elevator system, including structural parameters based on car mass, counterweight mass and wire rope parameters, motor and transmission system parameters, geometry and material parameters of guide rails and guide shoes, and elevator shaft environment parameters; Step S12: collecting the position, speed, acceleration, electrical parameters, vibration level and load changes during the operation of the elevator, so as to obtain the historical operation data of the elevator; Step S13: performing empirical mode decomposition on the elevator historical operation data and decomposing it into components of different time scales, thereby obtaining a multi-scale decomposition data set, wherein the different time scales include high-frequency vibration, medium-frequency motion, and low-frequency trend; Step S14: establishing an initial theoretical dynamic model according to the basic data of the elevator system; performing dynamic characteristic analysis at each time scale according to the initial theoretical dynamic model and the multi-scale decomposition data set, thereby obtaining multi-scale dynamic characteristic data, wherein the dynamic characteristic analysis at each time scale includes the identification of characteristic parameters at different operation stages, the characteristic analysis of the change of the wire rope stiffness with position, and the evaluation of the change law of the friction force with speed and temperature; Step S15: performing a multi-scale stability analysis on the multi-scale dynamic characteristic data according to the historical operation data of the elevator, thereby obtaining multi-scale stability data.

3. The elevator dynamics model correction method according to claim 2, characterized in that: Step S15 includes the following steps: Step S151: constructing stability analysis models at different time scales based on the multi-scale dynamic characteristic data, wherein the stability analysis model includes a high-frequency vibration stability model, a medium-frequency motion stability model, and a low-frequency trend stability model; Step S152: performing parameter identification and calibration on the stability analysis model according to the position, speed and acceleration information in the historical operation data of the elevator, and evaluating the stability boundary of the system based on different time scales, thereby obtaining stability boundary data, wherein the evaluation based on different time scales includes calculating the characteristic value and Lyapunov exponent of the stability boundary of the system at each time scale; Step S153: identifying key parameters and operating conditions that may cause system instability in the stability analysis model, thereby obtaining key parameter data; Step S154: Perform time domain and frequency domain stability analysis based on key parameter data and stability boundary data, and use statistical methods to perform stability analysis on the elevator historical operation data to generate multi-scale stability data.

4. The elevator dynamics model correction method according to claim 3, characterized in that: Step S2 includes the following steps: Step S21: establishing a multi-subsystem model of the elevator system according to the basic data of the elevator system, wherein the multi-subsystem model includes a car subsystem, a counterweight subsystem, a wire rope subsystem, a drive subsystem and a guide subsystem; Step S22: constructing a coupling relationship matrix between subsystems reflecting the degree of mutual influence between different subsystems based on the multi-scale stability data and the multi-subsystem model; Step S23: using a global sensitivity analysis method to identify the inter-subsystem coupling factors that have the greatest impact on system stability in the coupling relationship matrix, thereby obtaining stability impact data; Step S24: performing parameter change tests on the stability influencing data based on the orthogonal experimental design method, analyzing the influence of the parameter changes on the system stability, and establishing a mapping relationship model between the subsystem coupling parameters and the system stability using support vector regression; Step S25: calculating the sensitivity coefficient of each subsystem coupling parameter to the system stability in the mapping relationship model, and ranking the subsystem coupling parameters in importance according to the sensitivity coefficients, thereby obtaining a list of key coupling parameters; Step S26: construct a multiscale coupling sensitivity model that reflects the impact of subsystem coupling on system stability at different time scales based on the key coupling parameter list, and perform uncertainty analysis and optimization on the multiscale coupling sensitivity model to generate multiscale coupling sensitivity data.

5. The elevator dynamics model correction method according to claim 4, characterized in that: Step S3 includes the following steps: Step S31: Analyze the multi-scale coupling sensitivity data using a clustering algorithm to identify and classify the main nonlinear characteristics in the system, thereby obtaining nonlinear characteristic classification data; Step S32: performing nonlinear function fitting and selection based on the nonlinear characteristics of the nonlinear characteristic classification data, thereby constructing a nonlinear mapping function set; Step S33: integrating the nonlinear mapping function set according to the preset basic dynamic equation of the elevator system, thereby constructing a global nonlinear dynamic model; Step S34: designing a parameter identification model based on a machine learning algorithm according to the global nonlinear dynamics model, and designing an adaptive learning rate strategy to optimize the convergence performance of the parameter identification model, thereby generating an adaptive parameter identification strategy; Step S35: performing time series analysis on the multi-scale stability data based on the historical operation data of the elevator, and using the sliding time window method to identify the time-varying characteristics of the system parameters, thereby obtaining the parameter time-varying characteristic data; Step S36: Design a time-varying parameter estimation algorithm based on Kalman filtering according to the parameter time-varying characteristic data, and use the time-varying parameter estimation algorithm to generate an update strategy for the parameter time-varying characteristic data, thereby obtaining a time-varying parameter update strategy, wherein the update strategy specifically includes a parameter framework update strategy, a parameter update frequency strategy, a parameter smooth transition strategy, and a parameter update constraint strategy.

6. The elevator dynamics model correction method according to claim 5, characterized in that: Step S34 includes the following steps: Step S341: performing parameter sensitivity analysis to determine the key parameters that have the greatest impact on the model behavior based on the global nonlinear dynamics model, thereby obtaining parameter sensitivity data; Step S342: transforming the global nonlinear dynamics model into a mathematical expression for parameter identification based on key parameters according to the parameter sensitivity data, thereby obtaining a parameter identification problem expression; Step S343: setting initial values ​​for the parameters to be identified according to the parameter identification problem expression and the preset prior knowledge, and performing recursive least squares algorithm design and extended Kalman filter algorithm design, thereby obtaining recursive least squares and extended Kalman filter algorithm data, wherein the recursive least squares algorithm design specifically includes forgetting factor determination and covariance matrix initialization, wherein the extended Kalman filter algorithm design specifically includes linearization of state equations and observation equations and setting of process noise and observation noise covariance matrices; Step S344: generating an algorithm switching strategy based on estimation error according to the recursive least squares and extended Kalman filter algorithm data, thereby obtaining algorithm switching strategy data; Step S345: setting a value range and limiting a change rate for the parameter to be identified according to the algorithm switching strategy and the parameter identification problem expression, and designing a parameter constraint processing mechanism by a penalty function method, thereby obtaining constraint processing mechanism data; Step S346: Design a parameter identification model based on the machine learning algorithm according to the algorithm switching strategy data, constraint processing mechanism data, and recursive least squares and extended Kalman filter algorithm data, and design an adaptive learning rate strategy to optimize the convergence performance of the parameter identification model, thereby generating an adaptive parameter identification strategy.

7. The elevator dynamics model correction method according to claim 6, characterized in that: Step S4 includes the following steps: Step S41: acquiring real-time status monitoring data of the elevator system by deploying a sensor network, and performing preprocessing on the real-time status monitoring data based on denoising, filtering and data standardization, thereby obtaining real-time status data; Step S42: using a machine learning method to establish a normal operation mode of the elevator system and a feature library of various abnormal states according to historical operation data, thereby obtaining an abnormal state feature library; Step S43: comparing the real-time status data with the abnormal status feature library, and identifying potential abnormal status, thereby generating abnormal warning signal data; Step S44: activating corresponding modules according to abnormal warning signal data and adaptive parameter identification strategy, and quickly estimating short-term dynamic changes of the system, thereby generating short-term dynamic correction parameters; Step S45: evaluating the long-term trend of the system based on the time-varying parameter update strategy, the abnormal warning signal data, and the multi-scale dynamic characteristic data, thereby generating a long-term trend correction parameter; Step S46: Perform weighted fusion of short-term dynamic correction parameters and long-term trend correction parameters based on system changes at multiple time scales to generate comprehensive dynamic correction parameter data.

8. The elevator dynamics model correction method according to claim 7, characterized in that: Step S5 includes the following steps: Step S51: applying the dynamic correction parameter data to the global nonlinear dynamic model, and updating the model parameters and the model structure, thereby generating an updated dynamic model; Step S52: using the updated dynamics model to predict the short-term behavior of the elevator system, thereby generating model prediction result data; Step S53: Calculate statistical indicators of the model prediction result data and the real-time status monitoring data, thereby generating model accuracy evaluation indicator data; Step S54: evaluating the effectiveness of correcting the updated dynamics model according to the preset performance threshold and model accuracy evaluation index data, thereby generating model correction effect data; Step S55: adjusting the learning rate and forgetting factor in the adaptive parameter identification strategy based on the model correction effect data, thereby generating a parameter identification optimization strategy; Step S56: structurally optimizing the updated dynamics model according to the parameter identification optimization strategy and the model correction effect data, thereby generating optimized model structural data.

9. The elevator dynamics model correction method according to claim 8, characterized in that: Step S6 includes the following steps: Step S61: using the empirical mode decomposition method to perform high-frequency vibration mode recognition in the elevator system on the multi-scale dynamic characteristic data, thereby obtaining high-frequency vibration mode recognition data; Step S62: integrating the high-frequency vibration pattern recognition data into the state space model based on the optimized model structure data, and constructing an extended state space model; Step S63: performing an extended state observer design according to the extended state space model, thereby obtaining an extended state observer design scheme, wherein the extended state observer design includes determining an observer gain matrix and processing nonlinear terms and uncertainties; Step S64: using an extended state observer design scheme to estimate the state that is difficult to measure directly, thereby obtaining high-frequency vibration estimation data; Step S65: performing multi-subsystem collaborative optimization according to the high-frequency vibration estimation data and the multi-scale coupling sensitivity data to generate coupling optimization parameters between subsystems; Step S66: Correcting the global nonlinear dynamics model according to the inter-subsystem coupling optimization parameters to obtain a corrected elevator dynamics model.

10. An elevator dynamics model correction system, characterized in that: Used to execute the elevator dynamics model correction method according to claim 1, the elevator dynamics model correction system comprises: The basic data and multi-scale analysis module is used to obtain the basic data of the elevator system and the historical operation data of the elevator; perform multi-scale dynamic characteristics analysis based on the basic data of the elevator system to generate multi-scale dynamic characteristics data; perform multi-scale stability analysis on the multi-scale dynamic characteristics data based on the historical operation data of the elevator to obtain multi-scale stability data; A coupling sensitivity analysis module is used to perform subsystem coupling sensitivity analysis on multi-scale stability data according to structural parameters in the basic data of the elevator system, thereby generating multi-scale coupling sensitivity data; The nonlinear model and parameter identification module is used to perform global nonlinear characteristic mapping processing based on multi-scale coupling sensitivity data to obtain a global nonlinear dynamic model; perform parameter identification strategy processing based on the global nonlinear dynamic model to generate an adaptive parameter identification strategy; perform time-varying characteristic identification processing on multi-scale stability data based on the historical operation data of the elevator to generate a time-varying parameter update strategy; The real-time monitoring and dynamic correction module is used to obtain real-time status monitoring data of the elevator system, identify abnormal status based on the real-time status monitoring data, and obtain abnormal warning signal data; use adaptive parameter identification strategy and time-varying parameter update strategy to perform multi-time scale dynamic correction processing on abnormal warning signal data to obtain dynamic correction parameter data; The model updating and optimization module is used to perform model updating according to the dynamic correction parameter data, evaluate the model accuracy, and generate model correction effect data; based on the model correction effect data, the model structure is re-optimized according to the adaptive parameter identification strategy to generate optimized model structure data; An extended observation and correction module is used to construct an extended state observer based on the high-frequency vibration mode according to the optimized model structure data, and use the extended state observer to estimate the state that is difficult to measure directly, so as to obtain high-frequency vibration estimation data; Multi-subsystem collaborative optimization is performed based on high-frequency vibration estimation data and multi-scale coupling sensitivity data to generate inter-subsystem coupling optimization parameters; the global nonlinear dynamic model is corrected based on the inter-subsystem coupling optimization parameters to obtain a corrected elevator dynamic model.

Images