Method and terminal for intelligent analysis and risk prediction of elevator safety data

Abstract

The present application belongs to the technical field of elevator safety monitoring, and discloses a method and a terminal for intelligent analysis of elevator safety data and risk prediction; comprising: establishing an ideal rigid body motion trajectory based on elevator physical topology constraints; obtaining the real-time running trajectory of the elevator, comparing the trajectory deviation with the ideal rigid body motion trajectory, performing inversion based on the trajectory deviation to obtain mechanical impedance degradation parameters containing damping coefficient variation and stiffness coefficient variation; transforming the continuous variation in the mechanical impedance degradation parameters to form abnormal potential features to be verified, performing compliance logic discrimination on the abnormal potential features to be verified to form a sub-health state determination conclusion; performing counterfactual reasoning under the conditions of no load and full load based on the sub-health state determination conclusion to obtain predicted deviation response values under each load scenario, further judging the predicted deviation response values to generate an elevator comprehensive risk prediction report, which significantly improves the accuracy of elevator risk prediction.

Description

Technical Field

[0001] This invention relates to the field of elevator safety monitoring technology, and more specifically, to a method and terminal for intelligent analysis of elevator safety data and risk prediction. Background Technology

[0002] As the core vertical transportation equipment in high-rise buildings, elevators have received increasing attention for their operational safety and reliability. Traditional scheduled maintenance, due to drawbacks such as delayed maintenance windows or over-maintenance, is gradually evolving towards a condition-based predictive maintenance model.

[0003] In existing technologies, elevator risk prediction typically employs a sensor data-driven approach. This involves using sensors to collect operational signals from components such as the traction machine, guide shoes, and door operator. Machine learning is then used to extract time-domain, frequency-domain, or time-frequency-domain features from these signals to identify abnormal data, thereby enabling elevator fault warnings and risk assessments. However, the operational safety status of an elevator is inherently deeply coupled with its load variables. The dynamic response characteristics of mechanical components differ significantly under different load conditions. Existing elevator risk prediction methods often rely on data analysis from a single operational scenario to generate risk assessments, without considering multi-scenario extrapolation under different load conditions. For example, a certain wear condition may exhibit stable performance during no-load operation but trigger vibration responses far exceeding safety boundaries under full-load conditions. Due to the lack of extrapolation under different load scenarios, the risk prediction results fail to reflect the true risk level of the elevator under extreme or occasional load conditions, resulting in insufficient comprehensiveness and robustness of the prediction conclusions.

[0004] In view of this, the present invention proposes a method and terminal for intelligent analysis and risk prediction of elevator safety data to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: A method for intelligent analysis and risk prediction of elevator safety data includes: S1. Establish the ideal rigid body motion trajectory based on the elevator's physical topology constraints; S2. Obtain the real-time running trajectory of the elevator, compare the trajectory deviation with the ideal rigid body motion trajectory, perform inversion based on the trajectory deviation, and obtain the mechanical impedance degradation parameters including the change in damping coefficient and the change in stiffness coefficient. S3. Transform the continuous changes in the mechanical impedance degradation parameters to form abnormal potential features to be verified. Perform compliance logic judgment on the abnormal potential features to be verified. Perform semantic annotation on the abnormal potential features judged as compliant to form a sub-health state judgment conclusion. S4. Based on the sub-health status determination conclusion, perform counterfactual deduction under no-load and full-load scenarios to obtain the prediction deviation response value under each load scenario. Further judge the prediction deviation response value and generate an elevator comprehensive risk prediction report.

[0006] Furthermore, methods for establishing ideal rigid body motion trajectories based on elevator physical topology constraints include: The elevator physical topology constraints are pre-constructed. These constraints include geometric topology constraints that define the topological connections and spatial transformations between physical components, dynamic coupling constraints that define the mechanical transmission rules between physical components, and boundary condition constraints that limit the extreme values ​​of the motion of physical components. The physical components include the traction machine, wire rope, car frame, car, and guide rails. Obtain the target operating curve parameters of the elevator, including rated operating speed, rated acceleration, and maximum allowable acceleration; based on spatial transformation relationships and mechanical transmission rules, establish a motion transmission path starting from the traction machine, passing through the wire rope and car frame to the car in sequence; The target running curve parameters are used as inputs to the traction machine and transmitted step by step along the motion transmission path. The elevator physical topology constraints are called to convert the motion state transmitted to each physical component. The theoretical displacement, theoretical velocity and theoretical acceleration at the car are arranged according to the sampling time points to generate the ideal rigid body motion trajectory.

[0007] Furthermore, methods for comparing trajectory deviations from ideal rigid body motion trajectories include: The car's displacement at each sampling time point within the current elevator operating cycle is collected and differentiated to obtain real-time velocity and real-time acceleration. The car's displacement, real-time velocity, and real-time acceleration are then aligned and integrated according to the sampling time point to form a real-time operating trajectory. By comparing the various deviations between the real-time running trajectory and the ideal rigid body motion trajectory at each time point, the trajectory deviation, including position deviation, velocity deviation and acceleration deviation, is obtained.

[0008] Furthermore, methods for performing inversion based on trajectory deviation include: Using trajectory deviation as the input and equivalent damping coefficient and equivalent stiffness coefficient as the parameters to be solved, an equivalent mechanical impedance equation is constructed; the trajectory deviation is solved point by point through the equivalent mechanical impedance equation to obtain the equivalent damping coefficient and equivalent stiffness coefficient at each time point. The initial equivalent damping coefficient and initial equivalent stiffness coefficient values ​​recorded at the beginning of elevator use are obtained. The deviations from the equivalent damping coefficient and equivalent stiffness coefficient at each time point are compared to obtain the mechanical impedance degradation parameters at each time point, which include the changes in damping coefficient and stiffness coefficient.

[0009] Furthermore, methods for converting continuously varying quantities in the mechanical impedance degradation parameters include: Set up a codebook vector space, which includes the correspondence between the continuous range of values ​​of the damping coefficient change and the stiffness coefficient change and the discrete codeword, as well as the correspondence between the discrete codeword and the physical degradation type. By discretizing the numerical fluctuations of each mechanical impedance degradation parameter in the codebook vector space, a potential anomaly feature to be verified, including damping degradation codewords and stiffness degradation codewords, is formed.

[0010] Furthermore, methods for compliance logic judgment of potential abnormal characteristics to be verified include: Construct a Boolean logic discriminant, which includes a physical boundary constraint chain for verifying the legality of the values ​​of the potential features of the anomaly to be verified in the codebook vector space, and a temporal logic constraint chain for verifying the compliance of the evolution of the potential features of the anomaly to be verified between adjacent time points. Determine whether each potential feature of an anomaly to be verified satisfies the Boolean logic discriminant. If it does, it is deemed compliant; otherwise, it is deemed non-compliant.

[0011] Furthermore, the methods for arriving at a conclusion regarding a sub-healthy state include: Collect car vibration vectors at various time points within the current elevator operating cycle; quantify and statistically analyze the dependence of compliant potential anomalies on the corresponding car vibration vectors at those time points; If the dependency strength is higher than the preset mutual information threshold, the corresponding abnormal potential feature is judged to have a false association and is removed; through the correspondence between discrete codewords and physical degradation types in the codebook vector space, the retained abnormal potential features are semantically annotated to generate a sub-health status judgment conclusion containing semantic annotations.

[0012] Furthermore, methods for performing counterfactual inferences under no-load and full-load scenarios include: Based on the read elevator car self-weight mass parameters and rated load mass parameters, equivalent car mass parameters are established for both unloaded and fully loaded scenarios. The mechanical impedance degradation parameters corresponding to the current time point in the sub-health state judgment conclusion are extracted, and the equivalent mechanical impedance equation is forward derived by using the car equivalent mass parameters in the no-load and full-load scenarios respectively to obtain the predicted deviation response values ​​for each load scenario; the predicted deviation response values ​​include position deviation response, velocity deviation response and acceleration deviation response.

[0013] Furthermore, methods for further judging the prediction deviation response value include: By comparing the degree of deviation of the position deviation response, velocity deviation response, and acceleration deviation response under the full-load scenario with that under the no-load scenario, the deterioration factor of each response is obtained; A threshold range is preset, including a mild deterioration threshold and a severe deterioration threshold. Each response deterioration factor is compared with the threshold range. If any response deterioration factor is greater than the severe deterioration threshold, the elevator is determined to be at a high-risk level. If all response deterioration factors are less than the mild deterioration threshold, the elevator is determined to be at a low-risk level. Otherwise, the elevator is determined to be at a medium-risk level. The physical degradation type and semantic annotations are extracted from the sub-health status assessment conclusion. Combined with the deviation response type corresponding to the maximum response deterioration multiple, risk source information is generated. The risk source information is integrated with the risk level to generate a comprehensive elevator risk prediction report.

[0014] Furthermore, a terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements a method for intelligent analysis of elevator safety data and risk prediction.

[0015] The technical effects and advantages of the method and terminal for intelligent analysis and risk prediction of elevator safety data of this invention are as follows: 1. This invention constructs physical topological constraints for the elevator, generates an ideal rigid body motion trajectory along the motion transmission path as a benchmark, and compares it point by point with the real-time running trajectory to obtain the trajectory deviation. This provides a clear physical orientation for subsequent degradation parameter inversion. By performing inversion on the trajectory deviation through the equivalent mechanical impedance equation, mechanical impedance degradation parameters are obtained. This achieves accurate quantification from macroscopic motion deviation to the degree of degradation of microscopic physical properties, providing a prerequisite for accurate identification of the degradation state of elevator mechanical components and laying a solid foundation for the entire risk prediction process. 2. By setting up a codebook vector space to discretely map mechanical impedance degradation parameters, potential abnormal features to be verified are formed. Boolean logic discriminant is used for compliance verification to eliminate physically impossible violations caused by sensor noise or instantaneous disturbances. Furthermore, spurious correlations are further stripped away through quantitative statistics and the dependence strength of car displacement, realizing the transformation from numerical degradation to credible diagnostic conclusions, providing a reliable basis for subsequent risk extrapolation. Based on the sub-health state determination conclusion, counterfactual extrapolation is performed under no-load and full-load scenarios. Combining the car's equivalent mass parameters under different load scenarios, the prediction deviation response value under each load condition is extrapolated, enabling a forward-looking extrapolation and quantitative comparison of the risk evolution law of elevators under different load conditions, accurately capturing the risk differences caused by mechanical component degradation under different load conditions. A comprehensive elevator risk prediction report, including risk level and degradation tracing, is generated by combining the physical degradation type and degraded component name in the sub-health state determination conclusion, significantly improving the comprehensiveness, accuracy, and engineering practicality of elevator risk prediction. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the principle of an intelligent analysis and risk prediction method for elevator safety data according to the present invention. Figure 2 This is a flowchart for determining whether the potential features of an anomaly to be verified are compliant according to the present invention; Figure 3 This is a flowchart illustrating the process of determining the risk level of an elevator according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1, please refer to Figure 1 , Figure 2 and Figure 3 As shown in this embodiment, a method for intelligent analysis and risk prediction of elevator safety data includes: S1. Establish the ideal rigid body motion trajectory based on the physical topology constraints of the elevator.

[0019] Methods for establishing ideal rigid body motion trajectories based on elevator physical topology constraints include: The elevator physical topology constraints are pre-constructed. These constraints include geometric topology constraints that define the topological connection and spatial transformation relationships between physical components, dynamic coupling constraints that define the mechanical transmission rules between physical components, and boundary condition constraints that limit the extreme values ​​of the motion of physical components. The physical components include the traction machine, wire rope, car frame, car, and guide rails.

[0020] Elevator physical topology constraints are used to organize the dispersed physical components (traction machine, wire rope, car frame, car, guide rail) in an elevator into a constraint system with clear rules according to their actual spatial connection relationships, mechanical interaction laws and motion boundary conditions.

[0021] Among them, the topological connection relationship in the geometric topological constraints clarifies the connection sequence and connection method between each physical component. For example, it includes: the traction machine is installed in the machine room and connected to the wire rope through the traction sheave groove; one end of the wire rope passes around the traction sheave and connects to the car frame, and the other end connects to the counterweight; the car frame serves as a load-bearing frame to support the car; the car contacts the guide rail through the guide shoe and runs along the direction of the guide rail; the guide rail is fixed to the hoistway wall and limits the movement trajectory of the car.

[0022] The spatial transformation relationships in geometric topology constraints clarify the coordinate transformation relationships between two adjacent physical components. For example, the rotational motion of the traction machine output shaft is transformed into the linear motion of the wire rope in the vertical direction, the vertical motion of the lower end of the wire rope is transformed into the vertical motion of the car frame, and the vertical motion of the car frame is transformed into the vertical motion of the car.

[0023] Dynamic coupling constraints are used to simplify the connection elements between physical components into spring-damped elements and define the rules for the transmission of force and motion between adjacent mass blocks. For example, the mechanical transmission rules include: Mass block simplification rule: The traction machine, wire rope, car frame and car are simplified into corresponding mass blocks.

[0024] The rules for elastic connection elements are as follows: The wire rope is simplified to a spring-damping element connecting the traction machine mass block and the car frame mass block (wherein, the equivalent stiffness coefficient of the wire rope is determined by the wire rope material, diameter, and effective length; the equivalent damping coefficient is determined by the internal friction characteristics and relative slippage characteristics between the strands of the wire rope). The shock-absorbing pad between the car frame and the car is simplified to a spring-damping element connecting the car frame mass block and the car mass block (wherein, the equivalent stiffness coefficient and equivalent damping coefficient of the shock-absorbing pad are determined by the properties of the rubber material).

[0025] Mechanical transmission rules: Based on the displacement and velocity of the upstream mass block and the force acting on the downstream mass block, calculate the acceleration, velocity, and displacement of the downstream mass block according to Newton's second law.

[0026] Boundary condition constraints, based on elevator design specifications and safety standards, define hard upper limits on the displacement, velocity, and acceleration of each physical component. These include, for example: the maximum permissible displacement of the car on the guide rails (e.g., 80m); the maximum permissible operating speed of the car (e.g., 2.5m / s); and the maximum permissible acceleration of the car (e.g., 1.2m / s²). 2 ).

[0027] Obtain the target operating curve parameters of the elevator, including its rated operating speed, rated acceleration, and maximum allowable acceleration. The target operating curve parameters represent the ideal speed variation of the elevator car along the guide rails during a complete elevator cycle. These parameters are retrieved from the elevator's manufacturer's technical documentation.

[0028] Based on spatial transformation relationships and mechanical transmission rules, a motion transmission path is established starting from the traction machine, passing through the wire rope and car frame to the car. The target running curve parameters are converted from the linear motion of the car along the guide rail direction to the rotational motion of the traction machine output end, which serves as the input for the starting point of the motion transmission path.

[0029] The motion transmission path is determined by the actual sequence of force and motion transmission: the rotational motion of the traction machine output shaft is transmitted to the traction sheave; the traction sheave drives the wire rope through the friction of the sheave groove; the wire rope converts the rotational motion at the end of the traction sheave into vertical linear motion, which is transmitted to the car frame; the car frame, through the shock-absorbing pads, carries the car and transmits the motion to the car. Thus, the motion transmission path is established as "traction machine - wire rope - car frame - car".

[0030] The target running curve parameters are used as inputs to the traction machine and transmitted step by step along the motion transmission path. The elevator physical topology constraints are called to convert the motion state transmitted to each physical component. The theoretical displacement, theoretical velocity and theoretical acceleration at the car are arranged according to the sampling time points to generate the ideal rigid body motion trajectory.

[0031] Based on the spatial transformation relationship (rotational motion to linear motion) and mechanical transmission rules (spring-damping characteristics of the wire rope) between the traction sheave and the wire rope, the displacement, velocity, and acceleration output by the wire rope are obtained; based on the spatial transformation relationship between the lower end of the wire rope and the car frame, coordinate conversion is performed, and based on the mechanical transmission rules of the connection between the wire rope and the car frame, the displacement, velocity, and acceleration of the car frame are obtained. The transmission from the car frame to the car is calculated by coordinate transformation based on the spatial transformation relationship of the damping pads, and the theoretical displacement, theoretical velocity, and theoretical acceleration of the car are obtained according to the mechanical transmission rules of the damping pads.

[0032] S2. Obtain the real-time running trajectory of the elevator, compare the trajectory deviation with the ideal rigid body motion trajectory, perform inversion based on the trajectory deviation, and obtain the mechanical impedance degradation parameters including the changes in damping coefficient and stiffness coefficient.

[0033] Physical degradation of elevator mechanical components (such as guide shoe wear, wire rope slack, and shock absorber aging) directly manifests as drift in the equivalent damping coefficient and equivalent stiffness coefficient. These changes are reflected in the deviations in the elevator's motion state during operation.

[0034] For example, wear of the guide shoe liner causes changes in the frictional damping characteristics between the guide rail and the guide shoe, and the equivalent damping coefficient deviates from the initial value; the change in the equivalent damping coefficient is mainly reflected in the smoothness of the car's running speed, which manifests as an increase in speed deviation; A decrease in the elastic modulus of the wire rope or aging of the damping pads weakens the system's elastic recovery capability, causing the equivalent stiffness coefficient to deviate from its initial value. The change in the equivalent stiffness coefficient is mainly reflected in the accuracy of the car's displacement response, manifested as an increase in positional deviation. The coupling change between the equivalent damping coefficient and the equivalent stiffness coefficient is reflected in the degree of fluctuation in the car's acceleration, manifested as an increase in acceleration deviation.

[0035] Methods for comparing trajectory deviations from ideal rigid body motion trajectories include: The car's displacement at each sampling time point within the current elevator operating cycle is collected and differentiated to obtain the real-time velocity and real-time acceleration. The car displacement represents the car's absolute position on the guide rails. The first-order differential of the car displacement at each sampling time point is performed to obtain the real-time velocity at that time point. The second-order differential of the car displacement sequence is performed to obtain the real-time acceleration at each sampling time point.

[0036] It should be explained that the current operating cycle refers to the time interval during which the elevator completes one full operation, specifically from the start of operation after the car doors have closed until the car reaches the target floor and the car doors are ready to open. The sampling period for the car's displacement is consistent with the sampling period used to generate the ideal rigid body motion trajectory.

[0037] The car's displacement, real-time velocity, and real-time acceleration are aligned and integrated according to the sampling time point to form a real-time running trajectory. The car's displacement, real-time velocity, and real-time acceleration corresponding to the same sampling time point are paired and integrated to form a real-time motion state vector for that sampling time point. All real-time motion state vectors from the sampling time points are arranged in chronological order to form a real-time running trajectory sequence.

[0038] Under the conditions of the same sampling period and the same starting time, the real-time running trajectory and the ideal rigid body motion trajectory have the same time index and the same sequence length.

[0039] The trajectory deviation is obtained by comparing the deviations between the real-time running trajectory and the ideal rigid body motion trajectory at each time point, including position deviation, velocity deviation, and acceleration deviation. At each sampling time point, the position deviation is obtained by subtracting the theoretical displacement of the ideal rigid body motion trajectory from the car's running displacement in the real-time running trajectory; the velocity deviation is obtained by subtracting the theoretical velocity of the ideal rigid body motion trajectory from the real-time velocity in the real-time running trajectory; and the acceleration deviation is obtained by subtracting the theoretical acceleration of the ideal rigid body motion trajectory from the real-time acceleration in the real-time running trajectory. The position deviation, velocity deviation, and acceleration deviation at each sampling time point are summarized to form the trajectory deviation.

[0040] Methods for performing inversion based on trajectory deviation include: Using trajectory deviation as the input and equivalent damping coefficient and equivalent stiffness coefficient as the parameters to be solved, an equivalent mechanical impedance equation is constructed.

[0041] The equivalent mechanical impedance equation is expressed as follows: ;

[0042] in, This is the equivalent mass parameter; it represents the inertial characteristics of the elevator's moving parts. The equivalent mass parameter can be the sum of half the car's self-weight parameter and the rated load mass parameter (i.e., the intermediate representative value between the two extreme conditions of no-load and full-load). The acceleration deviation at the t-th sampling time point; This represents the velocity deviation at the t-th sampling time point; This is expressed as the positional deviation at the t-th sampling time point; This represents the equivalent damping coefficient, indicating the equivalent damping characteristics of the elevator. This represents the equivalent stiffness coefficient.

[0043] This represents the deviation of the external excitation force at the t-th sampling time point. During normal elevator operation, the driving force output by the traction machine is the same as under ideal conditions. There is no additional external excitation force from outside the traction machine during the car's movement along the guide rails; therefore, the external excitation force deviation is... Set to zero.

[0044] The trajectory deviation is solved point-by-point using the equivalent mechanical impedance equation to obtain the equivalent damping coefficient and equivalent stiffness coefficient at each time point. The solution method is as follows: A sliding window containing multiple consecutive sampling time points (e.g., 20 sampling time points) is defined. The position deviation, velocity deviation, and acceleration deviation of all sampling time points within the sliding window are substituted into the equivalent mechanical impedance equation to form an overdetermined system of equations regarding the equivalent damping coefficient and equivalent stiffness coefficient. The least squares method is applied to solve the overdetermined system of equations to obtain the estimated values ​​of the equivalent damping coefficient and equivalent stiffness coefficient corresponding to the sliding window. The sliding window is then traversed point by point along the sampling time points, and a new set of estimates is obtained with each traversal, until all sampling time points have been traversed, yielding the equivalent damping coefficient and equivalent stiffness coefficient at each sampling time point.

[0045] Obtain the initial equivalent damping coefficient and initial equivalent stiffness coefficient values ​​recorded during the initial period of elevator operation. The initial equivalent damping coefficient and initial equivalent stiffness coefficient values ​​are obtained by performing a complete cycle of operation when the mechanical components are in good condition during the initial period of elevator operation.

[0046] By comparing the deviations from the equivalent damping coefficient and equivalent stiffness coefficient at each time point, mechanical impedance degradation parameters containing the changes in damping coefficient and stiffness coefficient at each time point are obtained.

[0047] At each sampling time point, the difference between the equivalent damping coefficient and the initial equivalent damping coefficient is calculated to obtain the change in damping coefficient at each sampling time point. Similarly, the difference between the equivalent stiffness coefficient and the initial equivalent stiffness coefficient is calculated to obtain the change in stiffness coefficient at each sampling time point.

[0048] In this context, a positive change in the damping coefficient indicates an increase in the equivalent damping coefficient compared to the initial state, while a negative change indicates a decrease. Similarly, a positive change in the stiffness coefficient indicates an increase in the equivalent stiffness coefficient compared to the initial state, while a negative change indicates a decrease. An absolute value of zero indicates no deviation from the initial state.

[0049] S3. Transform the continuous changes in the mechanical impedance degradation parameters to form abnormal potential features to be verified. Perform compliance logic judgment on the abnormal potential features to be verified. Perform semantic annotation on the abnormal potential features judged as compliant to form a sub-health state judgment conclusion.

[0050] Methods for converting continuous variations in mechanical impedance degradation parameters include: A codebook vector space is set up, which includes the correspondence between the continuous range of values ​​of the damping coefficient change and the stiffness coefficient change and the discrete codeword, as well as the correspondence between the discrete codeword and the physical degradation type.

[0051] The codebook vector space divides the continuous range of values ​​for the changes in damping coefficient and stiffness coefficient into several discrete intervals, assigning a unique codeword identifier to each interval. The division is based on the percentage change rate of the damping coefficient relative to the initial equivalent damping coefficient value, and the percentage change rate of the stiffness coefficient relative to the initial equivalent stiffness coefficient value. Different interval boundary values ​​are used for the damping and stiffness dimensions due to their different physical characteristics. Each numerical interval corresponds to a discrete codeword, which is composed of a combination of the type of degenerate physical quantity and the degree of degradation.

[0052] For example, if the change in damping coefficient is less than ±10% of the initial equivalent damping coefficient value, the damping degradation code is "damping not degraded"; if the change in damping coefficient is between ±10% and ±30% of the initial equivalent damping coefficient value, the damping degradation code is "slight damping degradation"; if the change in damping coefficient is between ±30% and ±60% of the initial equivalent damping coefficient value, the damping degradation code is "moderate damping degradation"; if the change in damping coefficient is between ±60% and ±80% of the initial equivalent damping coefficient value, the damping degradation code is "severe damping degradation"; if the change in damping coefficient exceeds ±80% of the initial equivalent damping coefficient value, the damping degradation code is "severe damping degradation".

[0053] For example, if the change in stiffness coefficient is less than ±10% of the initial equivalent stiffness coefficient value, the stiffness degradation code is "no stiffness degradation"; if the change in stiffness coefficient is between ±10% and ±25% of the initial equivalent stiffness coefficient value, the stiffness degradation code is "slight stiffness degradation"; if the change in stiffness coefficient is between ±25% and ±50% of the initial equivalent stiffness coefficient value, the stiffness degradation code is "moderate stiffness degradation"; if the change in stiffness coefficient is between ±50% and ±70% of the initial equivalent stiffness coefficient value, the stiffness degradation code is "severe stiffness degradation"; and if the change in stiffness coefficient exceeds ±70% of the initial equivalent stiffness coefficient value, the stiffness degradation code is "severe stiffness degradation".

[0054] Physical degradation types describe the specific degradation patterns of elevator mechanical components, including friction-dominated degradation, elasticity-dominated degradation, and balanced degradation. The specific correspondences are as follows: The lowest level is defined as no degradation, and the highest level is defined as severe degradation. When the damping degradation codeword level is higher than the stiffness degradation codeword level at the same sampling time point, it corresponds to friction-dominated degradation, which physically means that the damping characteristics of the guide shoe and guide rail friction pair are dominated by degradation. When the stiffness degradation codeword level is higher than the damping degradation codeword level, it corresponds to elasticity-dominated degradation, which physically means that the stiffness characteristics of elastic elements such as wire ropes or shock absorbers are dominated by degradation. When the damping degradation codeword level and the stiffness degradation codeword level are the same, it corresponds to balanced degradation, which physically means that damping and stiffness characteristics degrade simultaneously.

[0055] By discretizing the numerical fluctuations of each mechanical impedance degradation parameter in the codebook vector space, a potential anomaly feature to be verified, including damping degradation codewords and stiffness degradation codewords, is formed.

[0056] Methods for performing compliance logic checks on potential abnormal characteristics to be verified include: Construct a Boolean logic discriminant, which includes a physical boundary constraint chain for verifying the legality of the values ​​of the potential features of the anomaly to be verified in the codebook vector space, and a temporal logic constraint chain for verifying the compliance of the evolution of the potential features of the anomaly to be verified between adjacent time points.

[0057] Boolean logic discriminants are used to determine whether codeword combinations in potential abnormal features conform to physically possible degradation and temporal evolution patterns.

[0058] The specific verification conditions for the physical boundary constraint chain are as follows: First, the levels of all degraded codewords exceed the highest severe degradation level defined in the codebook vector space. Second, the absolute value of the difference between the damping degradation codeword level and the stiffness degradation codeword level at the same sampling time point does not exceed the maximum allowed level difference (e.g., two levels) in the codebook vector space.

[0059] The output of the physical boundary constraint chain is true if both conditions are met; otherwise, the output is false.

[0060] The specific verification conditions for the timing logic constraint chain are as follows: First, the damping degradation codeword level and stiffness degradation codeword level at the current sampling time point are both no lower than the corresponding levels at the previous sampling time point; second, the difference in damping degradation codeword level and stiffness degradation codeword level between the current sampling time point and the previous sampling time point does not exceed one level. When both conditions are met simultaneously, the output of the timing logic constraint chain is true; if either condition is not met, the output is false.

[0061] Determine whether each potential feature of an anomaly to be verified satisfies the Boolean logic discriminant. If it does, it is deemed compliant; otherwise, it is deemed non-compliant.

[0062] The methods for determining a sub-healthy state include: The car vibration vector is collected at various time points within the current elevator operating cycle. The car vibration vector includes vibration acceleration components in three directions: horizontal longitudinal, horizontal lateral, and vertical. When the guide shoes wear, the wire rope loosens, or the shock-absorbing pads age, the vibration characteristics of the car during operation will change accordingly. Therefore, the car vibration vector can be used to verify the authenticity of potential abnormal characteristics.

[0063] The dependence strength between the potential abnormal features of compliance and the corresponding car vibration vector at each time point is quantified. The specific methods for obtaining the dependence strength include: treating the damping degradation codeword level and stiffness degradation codeword level in the potential compliance abnormal features as a set of discrete random variables, and treating the composite amplitude of the three components of the car vibration vector at the same sampling time point as a continuous random variable; statistically analyzing the joint distribution and marginal distribution of the two variables at all sampling time points within the current operating cycle, obtaining the mutual information value using the mutual information calculation method, and using this value as the dependence strength corresponding to the potential compliance abnormal feature.

[0064] If the dependency strength is higher than the preset mutual information threshold, the corresponding abnormal potential feature is judged to have a false association and is removed.

[0065] The mutual information threshold is set based on the elevator installation and acceptance and the mechanical components being in a healthy state. Specifically, it involves generating potential compliant anomalies in a healthy state from data generated during multiple normal operations, calculating the mutual information value between the data and the car vibration vector, and taking the statistical upper limit of the multiple calculation results as the mutual information threshold.

[0066] Among them, the mutual information value between a certain compliance anomaly potential feature and the car vibration vector is higher than the mutual information threshold, indicating that there is a strong statistical dependence between the compliance anomaly potential feature and the car vibration vector that exceeds the normal fluctuation range. This dependence is more likely to be caused by non-mechanical degradation factors such as sensor common-mode noise, power fluctuation or signal transmission interference, rather than a causal relationship caused by the actual degradation of non-mechanical components.

[0067] By leveraging the correspondence between discrete codewords and physical degradation types in the codebook vector space, semantic annotations are applied to the retained potential abnormal features, generating sub-health status judgment conclusions that include semantic annotations.

[0068] Extract the damping degradation codeword and stiffness degradation codeword from each retained potential feature of compliance anomalies. Based on the correspondence between discrete codewords and physical degradation types, determine the physical degradation type corresponding to the sampling time point. Combine the degradation type with the degradation severity level corresponding to the discrete codeword to generate the physical annotation for that sampling time point.

[0069] For example, the damping degradation code is "moderate damping degradation" and the stiffness degradation code is "slight stiffness degradation". The damping degradation level is higher than the stiffness degradation level, which corresponds to friction-dominated degradation. The generated physical annotation is "friction-dominated moderate degradation, which is characterized by a decrease in the damping characteristics of the friction pair between the guide shoe and the guide rail, accompanied by a slight degradation in stiffness characteristics".

[0070] S4. Based on the sub-health status determination conclusion, perform counterfactual deduction under no-load and full-load scenarios to obtain the prediction deviation response value under each load scenario. Further judge the prediction deviation response value and generate an elevator comprehensive risk prediction report.

[0071] Given that the elevator is currently in a sub-healthy state, we can predict the risk level that the elevator may encounter in the future by assuming different load scenarios and conducting forward extrapolation, thus providing a forward-looking basis for operation and maintenance decisions.

[0072] The no-load scenario corresponds to the elevator's operating state under no-load conditions, while the full-load scenario corresponds to the elevator's operating state under the maximum design load conditions. These two scenarios represent the lower and upper limits of the elevator's load range, respectively. By comparing the predicted deviation response values ​​and risk levels under these two extreme load scenarios, the sensitivity of the current sub-optimal state to load changes can be obtained, providing a more comprehensive basis for risk level determination.

[0073] Methods for performing counterfactual inferences under no-load and full-load scenarios include: Based on the read elevator car self-weight mass parameters and rated load mass parameters, equivalent mass parameters of the car are established for both unloaded and fully loaded scenarios. The self-weight mass parameter represents the structural mass of the car itself; the rated load mass parameter represents the maximum allowable load capacity of the elevator design.

[0074] In the empty-load scenario, the car is set to have no passengers or cargo, and the car's equivalent mass parameter equals its own weight parameter. In the fully loaded scenario, the car is set to carry the rated load mass, and the car's equivalent mass parameter equals its own weight mass parameter plus the rated load mass parameter.

[0075] The mechanical impedance degradation parameters corresponding to the current time point in the sub-health state judgment conclusion are extracted, and the equivalent mechanical impedance equation is forward derived by using the car equivalent mass parameters in the no-load and full-load scenarios respectively to obtain the predicted deviation response values ​​for each load scenario; the predicted deviation response values ​​include position deviation response, velocity deviation response and acceleration deviation response.

[0076] Among them, the position deviation response represents the predicted deviation of the car's actual displacement from the ideal displacement under the current sub-healthy state and corresponding load scenario, reflecting the deteriorating trend of leveling accuracy. The speed deviation response represents the predicted deviation of the car's actual speed from the ideal speed, reflecting the deteriorating trend of operational smoothness. The acceleration deviation response represents the predicted deviation of the car's actual acceleration from the ideal acceleration, reflecting the deteriorating trend of operational impact.

[0077] Further methods for judging the prediction deviation response value include: By comparing the deviations in position, velocity, and acceleration responses under the full-load scenario compared to the no-load scenario, the degradation factor of each response is obtained. The degradation factor measures the degree of degradation of each deviation response under the full-load scenario compared to the no-load scenario.

[0078] The deterioration factor of the position deviation response = position deviation response under full load ÷ position deviation response under no-load condition; the larger the position deviation response deterioration factor, the more significant the impact of increased load on the leveling accuracy of the car. The deterioration factor of the speed deviation response = speed deviation response under full load ÷ speed deviation response under no-load condition; the larger the speed deviation response deterioration factor, the more significant the impact of increased load on the smoothness of car operation. The deterioration factor of the acceleration deviation response = acceleration deviation response under full load ÷ acceleration deviation response under no-load condition; the larger the acceleration deviation response deterioration factor, the more significant the impact of increased load on the impact of increased load on the car's running shock.

[0079] The system presets threshold ranges for mild and severe deterioration. These thresholds serve as the criterion for distinguishing different risk levels. The mild and severe deterioration thresholds are determined based on elevator type, rated operating parameters, and historical operating data. For example, a mild deterioration threshold of 2 means that a full-load deviation not exceeding twice the no-load deviation is considered a slight deterioration; a severe deterioration threshold of 5 means that a full-load deviation exceeding five times the no-load deviation is considered a serious deterioration.

[0080] By comparing each response degradation factor with a threshold range, if any response degradation factor exceeds the severe degradation threshold, the elevator is determined to be at a high-risk level. This indicates that the deviation response has severely deteriorated under full-load conditions, and the current sub-optimal state may lead to significantly unsafe operational deviations under full-load conditions.

[0081] If all response degradation factors are less than the mild degradation threshold, the elevator is considered to be at a low-risk level. This indicates that even under full load, the degradation of each deviation response is within a slight range, and the overall safe operation capability of the elevator is not substantially affected.

[0082] Otherwise, the elevator is classified as medium-risk. That is, at least one response degradation factor is between the mild degradation threshold and the severe degradation threshold, but no response degradation factor exceeds the severe degradation threshold; this indicates that the current sub-healthy state has a perceptible impact on the operating performance under full load conditions, but has not yet reached a serious level, and preventive maintenance is recommended.

[0083] The physical degradation type and semantic annotations are extracted from the sub-health status assessment conclusions. Combined with the deviation response type corresponding to the maximum response deterioration multiple, risk tracing information is generated. This risk tracing information is then integrated with the risk level to generate a comprehensive elevator risk prediction report.

[0084] The maximum response degradation factor refers to the highest value among the position deviation response degradation factor, velocity deviation response degradation factor, and acceleration response deviation degradation factor, reflecting the most prominent deterioration dimension of the current sub-healthy state under full load conditions. Extracting the deviation response type corresponding to the maximum response degradation factor indicates the motion dimension where the current sub-healthy state first exposes problems during full load operation, i.e., "what are the most important safety indicators to focus on under full load conditions?"

[0085] In this embodiment, by constructing physical topological constraints for the elevator, an ideal rigid body motion trajectory is generated along the motion transmission path as a benchmark, and the trajectory deviation is obtained by comparing it point by point with the real-time running trajectory. This provides a clear physical orientation for subsequent degradation parameter inversion. The trajectory deviation is inverted through the equivalent mechanical impedance equation to obtain the mechanical impedance degradation parameters. This achieves accurate quantification from macroscopic motion deviation to the degree of degradation of microscopic physical properties, providing a prerequisite for accurate identification of the degradation state of elevator mechanical components and laying a solid foundation for the entire risk prediction process.

[0086] By setting up a codebook vector space to discretely map mechanical impedance degradation parameters, potential abnormal features to be verified are formed. Boolean logic discriminant is used for compliance verification, eliminating physically impossible violations caused by sensor noise or instantaneous disturbances. Furthermore, spurious correlations are further stripped away through quantitative statistics and the dependence strength of car displacement, realizing the transformation from numerical degradation to credible diagnostic conclusions, providing a reliable basis for subsequent risk extrapolation. Based on the sub-health state determination conclusion, counterfactual extrapolation is performed under no-load and full-load scenarios. Combining the car's equivalent mass parameters under different load scenarios, the prediction deviation response value under each load condition is extrapolated, enabling a forward-looking extrapolation and quantitative comparison of the risk evolution law of elevators under different load conditions, accurately capturing the risk differences caused by mechanical component degradation under different load conditions. Finally, by combining the physical degradation type and degraded component name in the sub-health state determination conclusion, a comprehensive elevator risk prediction report including risk level and degradation tracing is generated, significantly improving the comprehensiveness, accuracy, and engineering practicality of elevator risk prediction.

[0087] Example 2: A terminal described in this example includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements a method for intelligent analysis and risk prediction of elevator safety data.

[0088] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0089] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of units is only one method, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0090] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

[0091] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent analysis and risk prediction of elevator safety data, characterized in that, include: S1. Establish the ideal rigid body motion trajectory based on the elevator's physical topology constraints; S2. Obtain the real-time running trajectory of the elevator, compare the trajectory deviation with the ideal rigid body motion trajectory, perform inversion based on the trajectory deviation, and obtain the mechanical impedance degradation parameters including the change in damping coefficient and the change in stiffness coefficient. S3. Transform the continuous changes in the mechanical impedance degradation parameters to form abnormal potential features to be verified. Perform compliance logic judgment on the abnormal potential features to be verified. Perform semantic annotation on the abnormal potential features judged as compliant to form a sub-health state judgment conclusion. S4. Based on the sub-health status determination conclusion, perform counterfactual deduction under no-load and full-load scenarios to obtain the prediction deviation response value under each load scenario. Further judge the prediction deviation response value and generate an elevator comprehensive risk prediction report.

2. The method for intelligent analysis and risk prediction of elevator safety data according to claim 1, characterized in that, The method for establishing an ideal rigid body motion trajectory based on elevator physical topology constraints includes: The elevator physical topology constraints are pre-constructed. These constraints include geometric topology constraints that define the topological connections and spatial transformations between physical components, dynamic coupling constraints that define the mechanical transmission rules between physical components, and boundary condition constraints that limit the extreme values ​​of the motion of physical components. The physical components include the traction machine, wire rope, car frame, car, and guide rails. Obtain the target operating curve parameters of the elevator, including rated operating speed, rated acceleration, and maximum allowable acceleration; based on spatial transformation relationships and mechanical transmission rules, establish a motion transmission path starting from the traction machine, passing through the wire rope and car frame to the car in sequence; The target running curve parameters are used as inputs to the traction machine and transmitted step by step along the motion transmission path. The elevator physical topology constraints are called to convert the motion state transmitted to each physical component. The theoretical displacement, theoretical velocity and theoretical acceleration at the car are arranged according to the sampling time points to generate the ideal rigid body motion trajectory.

3. The method for intelligent analysis and risk prediction of elevator safety data according to claim 2, characterized in that, The method for comparing the trajectory deviation with the ideal rigid body motion trajectory includes: The car's displacement at each sampling time point within the current elevator operating cycle is collected and differentiated to obtain real-time velocity and real-time acceleration. The car's displacement, real-time velocity, and real-time acceleration are then aligned and integrated according to the sampling time point to form a real-time operating trajectory. By comparing the various deviations between the real-time running trajectory and the ideal rigid body motion trajectory at each time point, the trajectory deviation, including position deviation, velocity deviation and acceleration deviation, is obtained.

4. The method for intelligent analysis and risk prediction of elevator safety data according to claim 3, characterized in that, The method for performing inversion based on trajectory deviation includes: Using trajectory deviation as the input and equivalent damping coefficient and equivalent stiffness coefficient as the parameters to be solved, an equivalent mechanical impedance equation is constructed; the trajectory deviation is solved point by point through the equivalent mechanical impedance equation to obtain the equivalent damping coefficient and equivalent stiffness coefficient at each time point. The initial equivalent damping coefficient and initial equivalent stiffness coefficient values ​​recorded at the beginning of elevator use are obtained. The deviations from the equivalent damping coefficient and equivalent stiffness coefficient at each time point are compared to obtain the mechanical impedance degradation parameters at each time point, which include the changes in damping coefficient and stiffness coefficient.

5. The method for intelligent analysis and risk prediction of elevator safety data according to claim 4, characterized in that, The method for converting the continuous variation in the mechanical impedance degradation parameter includes: Set up a codebook vector space, which includes the correspondence between the continuous range of values ​​of the damping coefficient change and the stiffness coefficient change and the discrete codeword, as well as the correspondence between the discrete codeword and the physical degradation type. By discretizing the numerical fluctuations of each mechanical impedance degradation parameter in the codebook vector space, a potential anomaly feature to be verified, including damping degradation codewords and stiffness degradation codewords, is formed.

6. The method for intelligent analysis and risk prediction of elevator safety data according to claim 5, characterized in that, The method for performing compliance logic judgment on the potential abnormal features to be verified includes: Construct a Boolean logic discriminant, which includes a physical boundary constraint chain for verifying the legality of the values ​​of the potential features of the anomaly to be verified in the codebook vector space, and a temporal logic constraint chain for verifying the compliance of the evolution of the potential features of the anomaly to be verified between adjacent time points. Determine whether each potential feature of an anomaly to be verified satisfies the Boolean logic discriminant. If it does, it is deemed compliant; otherwise, it is deemed non-compliant.

7. The method for intelligent analysis and risk prediction of elevator safety data according to claim 6, characterized in that, The methods for forming a sub-health state assessment conclusion include: Collect car vibration vectors at various time points within the current elevator operating cycle; quantify and statistically analyze the dependence of compliant potential anomalies on the corresponding car vibration vectors at those time points; If the dependency strength is higher than the preset mutual information threshold, the corresponding abnormal potential feature is judged to have a false association and is removed; through the correspondence between discrete codewords and physical degradation types in the codebook vector space, the retained abnormal potential features are semantically annotated to generate a sub-health status judgment conclusion containing semantic annotations.

8. The method for intelligent analysis and risk prediction of elevator safety data according to claim 7, characterized in that, The method for performing counterfactual inferences under no-load and full-load scenarios includes: Based on the read elevator car self-weight mass parameters and rated load mass parameters, equivalent car mass parameters are established for both unloaded and fully loaded scenarios. The mechanical impedance degradation parameters corresponding to the current time point in the sub-health state judgment conclusion are extracted, and the equivalent mechanical impedance equation is forward derived by using the car equivalent mass parameters in the no-load and full-load scenarios respectively to obtain the predicted deviation response values ​​for each load scenario; the predicted deviation response values ​​include position deviation response, velocity deviation response and acceleration deviation response.

9. The method for intelligent analysis and risk prediction of elevator safety data according to claim 8, characterized in that, The method for further determining the prediction deviation response value includes: By comparing the degree of deviation of the position deviation response, velocity deviation response, and acceleration deviation response under the full-load scenario with that under the no-load scenario, the deterioration factor of each response is obtained; A threshold range is preset, including a mild deterioration threshold and a severe deterioration threshold. Each response deterioration factor is compared with the threshold range. If any response deterioration factor is greater than the severe deterioration threshold, the elevator is determined to be at a high-risk level. If all response deterioration factors are less than the mild deterioration threshold, the elevator is determined to be at a low-risk level. Otherwise, the elevator is determined to be at a medium-risk level. The physical degradation type and semantic annotations are extracted from the sub-health status assessment conclusion. Combined with the deviation response type corresponding to the maximum response deterioration multiple, risk source information is generated. The risk source information is integrated with the risk level to generate a comprehensive elevator risk prediction report.

10. A terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements a method for intelligent analysis and risk prediction of elevator safety data as described in any one of claims 1 to 9.

Images