Elevator operation and maintenance full-life-cycle fault monitoring and diagnosing method and system

Abstract

The invention discloses an elevator operation and maintenance full-life-cycle fault monitoring and diagnosing method and system, and relates to the technical field of data analysis, and the method comprises the following steps: collecting fault action data of known elevator component parts, and establishing an elevator component part fault tree map; marking a historical operation process instruction of the target elevator equipment, analyzing the matching between the historical operation process instruction and the assembled building, and evaluating the current state of each component of the target elevator equipment; running data of all parts of the real-time target elevator equipment are obtained, similar fault screening is conducted on the running data and the known elevator component fault tree atlas, an undetermined fault set of all the parts of the real-time target elevator equipment is obtained, an elevator part-fault confidence coefficient function is established, and fault types of all the parts of the target elevator equipment are generated. The method has the advantages that the fault early recognition capability and the fault diagnosis accuracy are improved, the fault processing time is shortened, and the reliability of elevator equipment is improved.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a method and system for fault monitoring and diagnosis throughout the entire lifecycle of elevator operation and maintenance. Background Technology

[0002] Existing elevator maintenance fault monitoring and diagnosis systems suffer from severe mismatches between elevator assembly and building use. Standardized elevator designs fail to meet the specific needs of buildings (e.g., hospital overload, high-rise building pressure changes), the installation environment does not consider building structural characteristics (e.g., shaft resonance, coastal corrosion), diagnostic models are detached from actual complex operating conditions (e.g., high-frequency use of logistics elevators), and monitoring strategies are not updated synchronously after building function changes (e.g., office buildings converted into kindergartens). This results in traditional monitoring systems based on ideal operating conditions having high false alarm and false negative rates in practical applications, and failing to provide early warnings of potential faults caused by building characteristics. Summary of the Invention

[0003] To address the aforementioned technical problems, this paper provides a method and system for monitoring and diagnosing elevator operation and maintenance throughout its entire lifecycle. This solution resolves the issues arising from existing elevator operation and maintenance fault monitoring and diagnosis systems. These problems stem from a severe mismatch between elevator assembly and building use, resulting in standardized elevator designs failing to adapt to specific building needs (such as hospital overload or high-rise building pressure changes), installation environments not considering building structural characteristics (such as shaft resonance or coastal corrosion), diagnostic models detached from actual complex operating conditions (such as high-frequency use of logistics elevators), and monitoring strategies not being updated synchronously after building function changes (such as converting an office building into a kindergarten). These issues lead to high false alarm and false negative rates in traditional monitoring systems based on ideal operating conditions, and an inability to predict potential faults caused by building characteristics in practical applications.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] A method for fault monitoring and diagnosis throughout the entire lifecycle of elevator operation and maintenance includes:

[0006] S1. Based on the Internet, collect known elevator component failure action data, analyze the correlation trend between elevator component failure action data, and establish elevator component fault tree map.

[0007] S2. Based on the historical operation log of the target elevator equipment, mark the historical operation process instructions of the target elevator equipment, analyze the matching between the historical operation process instructions and the assembled building, and evaluate the current status of each component of the target elevator equipment;

[0008] S3. Based on edge sensors, obtain the real-time operating data of each component of the target elevator equipment and perform similar fault screening with the known elevator component fault tree map to obtain the set of undetermined faults for each component of the real-time target elevator equipment.

[0009] S4. Based on the current status of each component of the target elevator equipment and the set of undetermined faults of each component of the target elevator equipment, establish an elevator component-fault confidence function to generate the fault types of each component of the target elevator equipment.

[0010] Preferably, step S1 includes the following:

[0011] Based on the elevator equipment manufacturer's fault database and elevator equipment fault case library, obtain known elevator component fault action data;

[0012] Based on known elevator component failure action data, NLP (Natural Language Processing) is used to extract the causal relationship between the failure phenomenon and the failure cause of the elevator component. The failure, failure phenomenon and failure cause of each elevator component are used as independent nodes to build a known elevator component failure action connection matrix.

[0013] Based on the transfer entropy, calculate the causal intensity entropy value between each pair of nodes in the known elevator component fault action connection matrix, and determine the causal chain between each pair of nodes.

[0014] Based on the known fault action linkage matrix of elevator components and the causal chain between each pair of nodes, a weighted directed graph of the causal network of known elevator component fault actions is constructed.

[0015] Preferably, step S1 further includes:

[0016] Based on Node2Vec random walk machine learning, an initial vector is assigned to each node in the weighted directed graph of the causal network of known elevator component failure actions, and the maximum random walk transition probability between the initial vectors of each node is calculated to generate a causal neighbor sequence of known elevator component failure actions.

[0017] Using the HDBSCAN clustering algorithm, we verify the prior knowledge between nodes in the causal neighbor sequence of known elevator component failure actions, determine the neighborhood radius and minimum density number to divide the nodes, generate the minimum spanning tree of the causal reachability metric of failure actions, and extract the causal association clusters of known elevator component failure actions.

[0018] Based on the known causal relationship clusters of elevator component failure actions, calculate the inter-cluster relationship with the largest causal flow between each pair of clusters as parent and child nodes, and build a fault tree diagram of the known elevator components.

[0019] Preferably, step S2 includes the following:

[0020] Based on the historical operation logs of the target elevator equipment and the prefabricated building, extract the historical operation process instruction feature data of the target elevator equipment and the usage attribute feature data of the prefabricated building; the historical operation process instruction feature data includes: operation direction and operation speed, while the usage attributes of the prefabricated building include: building type, building functional zoning and building personnel flow.

[0021] Based on sliding window statistics, the runtime of the historical operation process instructions of the target elevator equipment is used as the observation window, and the unit time is used as the observation step. The historical operation process instruction feature data of the target elevator equipment is used as the observation object. The mean and variance of the historical operation process instructions within the sliding window are calculated to obtain the load status label of the historical operation process instructions of the target elevator equipment.

[0022] Based on principal component analysis, weights are assigned to the usage attribute feature data of prefabricated buildings, and the usage attribute feature vector index of prefabricated buildings is calculated.

[0023] Preferably, step S2 further includes:

[0024] The historical operation process instruction load status labels of the target elevator equipment and the feature vector index of the building's purpose attribute are normalized.

[0025] Based on DTW dynamic time warping, the cosine similarity between the historical operation process instruction load status label of the target elevator equipment and the feature vector index of the purpose attribute of the prefabricated building under a unit time is calculated to determine the matching degree between the historical operation process instruction of the target elevator equipment and the prefabricated building under a unit time.

[0026] Obtain historical standardized operating data of similar target elevator equipment in known prefabricated buildings.

[0027] Preferably, step S2 further includes:

[0028] Based on the historical standardized operation data of known prefabricated buildings and similar target elevator equipment, a GRU neural network is trained to establish a life prediction loss function for known prefabricated buildings and similar target elevator equipment. The historical operation process instruction feature data of the target elevator equipment and the usage attribute feature data of the prefabricated building are used as inputs, and the remaining health indicators of the target elevator equipment are used as outputs.

[0029] By using the historical operating process instructions of the target elevator equipment per unit time and the assembly building matching degree, the remaining health indicators of the target elevator equipment are corrected to obtain the health status indicators of each component of the target elevator equipment.

[0030] Preferably, step S3 includes the following:

[0031] Based on the real-time operational data of each component of the target elevator equipment, multi-dimensional characteristic parameters of each component of the real-time target elevator equipment are marked; the multi-dimensional characteristic parameters include: vibration parameters, temperature parameters, and current and voltage parameters;

[0032] Based on the design service life and basic performance degradation coefficient of each component of the real-time target elevator equipment and the cumulative operating time of each component of the real-time target elevator equipment, the dynamic fault threshold of each component of the real-time target elevator equipment is estimated.

[0033] Using the analytic hierarchy process (AHP), weights are assigned to the multi-dimensional operational characteristic parameters of each component of the target elevator equipment. Based on the multi-dimensional operational characteristic parameters of each component of the target elevator equipment and the known fault tree diagram of the elevator components, the weighted fault similarity index of each component of the target elevator equipment is calculated.

[0034] The system selects fault actions whose weighted fault similarity index for each component of the target elevator equipment is greater than the dynamic fault threshold of each component of the target elevator equipment, and constructs a set of undetermined faults for each component of the target elevator equipment.

[0035] Preferably, step S4 includes the following:

[0036] Verify the support value of the operating data of each component of the target elevator equipment for each undetermined fault in the set of undetermined faults of each component of the target elevator equipment. Using DS evidence theory, fuse the combined probabilities of the support values ​​of each undetermined fault in the set of undetermined faults of each component of the target elevator equipment to obtain the joint probability of the operating data of each component of the target elevator equipment for each undetermined fault of the target elevator equipment.

[0037] Based on the joint probability of undetermined faults of each component of the target elevator equipment and the fault tree diagram of known elevator components, the prior probability of faults of each component of the target elevator equipment is obtained.

[0038] Based on the operating data of each component of the target elevator equipment and the set of undetermined faults of each component of the target elevator equipment, a likelihood function is established to verify the likelihood probability of the operating data of each component of the target elevator equipment under the undetermined faults of each component.

[0039] Based on the prior probability of each component failure of the target elevator equipment and the likelihood probability of the operating data of each component of the target elevator equipment under the condition of undetermined failure of each component, the posterior confidence of each component failure of the target elevator equipment is calculated.

[0040] Preferably, step S4 further includes:

[0041] The health status indicators of each component of the target elevator equipment are normalized and used as the weights of the health status indicators of each component of the target elevator equipment.

[0042] Based on the weights of the health status indicators of each component of the target elevator equipment, the posterior confidence of the faults of each component of the target elevator equipment, and the set of undetermined faults of each component of the target elevator equipment, an elevator component status-fault confidence function is established to calculate the confidence of the fault type of each component of the target elevator equipment, as follows:

[0043] S i =α·P(F i |S)+β·w i ·γ·Severity(F i )

[0044] Among them, S i Let P(F) be the confidence level of the fault type of the i-th component of the target elevator equipment. i |S) represents the posterior confidence level of the failure of the i-th component given the operating data of each component of the target elevator equipment, w i Severity(F) represents the weight of the remaining health indicators for the i-th component of the target elevator equipment. i Let α be the risk index of the failure of the i-th component of the target elevator equipment, β be the health weight coefficient, and γ be the failure risk coefficient.

[0045] Furthermore, an elevator operation and maintenance full lifecycle fault monitoring and diagnosis system is provided to implement the elevator operation and maintenance full lifecycle fault monitoring and diagnosis method described above, comprising:

[0046] Fault diagram module, elevator component estimation module, elevator fault screening module, and elevator fault location module;

[0047] The fault map module is used to collect known elevator component fault action data based on the Internet, analyze the correlation trends between elevator component fault action data, and establish elevator component fault tree map.

[0048] The elevator component estimation module is used to mark the historical operation process instructions of the target elevator equipment based on the historical operation log of the target elevator equipment, analyze the matching between the historical operation process instructions and the assembly building, and evaluate the current status of each component of the target elevator equipment.

[0049] The elevator fault screening module is electrically connected to the fault map module and the elevator component estimation module. The elevator fault screening module is used to obtain the real-time operating data of each component of the target elevator equipment based on edge sensors and perform similar fault screening with the known elevator component fault tree map to obtain the set of undetermined faults of each component of the real-time target elevator equipment.

[0050] The elevator fault location module is electrically connected to the elevator fault screening module. The elevator fault location module is used to establish an elevator component-fault confidence function based on the current status of each component of the target elevator equipment and the set of pending faults of each component of the target elevator equipment, and generate the fault types of each component of the target elevator equipment.

[0051] Compared with existing technologies, the advantages of this invention are as follows: This invention proposes a fault monitoring and diagnosis scheme for the entire lifecycle of elevator operation and maintenance. It constructs a fault tree graph by collecting historical fault data through the Industrial Internet, and assesses the health status of components by combining mining command matching analysis. It generates a set of undetermined faults by using real-time data from edge sensors and the fault graph for similarity matching. Finally, based on evidence theory, it integrates real-time monitoring data and equipment health indicators, and establishes a state-fault confidence function through Bayesian dynamic reasoning to achieve accurate fault type determination. The advantages are: improved early fault identification capability and fault diagnosis accuracy, reduced fault handling time, and improved reliability of elevator equipment. Attached Figure Description

[0052] Figure 1 A flowchart of a fault monitoring and diagnosis method for the entire life cycle of elevator operation and maintenance;

[0053] Figure 2 This is a framework diagram of a fault monitoring and diagnosis system for the entire life cycle of elevator operation and maintenance. Detailed Implementation

[0054] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0055] Reference Figure 1 As shown, a method for fault monitoring and diagnosis throughout the entire lifecycle of elevator operation and maintenance includes:

[0056] S1. Based on the Internet, collect known elevator component failure action data, analyze the correlation trend between elevator component failure action data, and establish elevator component fault tree map.

[0057] Step S1 includes the following:

[0058] Based on the elevator equipment manufacturer's fault database and elevator equipment fault case library, obtain known elevator component fault action data;

[0059] Based on known elevator component failure action data, NLP (Natural Language Processing) is used to extract the causal relationship between the failure phenomenon and the failure cause of the elevator component. The failure, failure phenomenon and failure cause of each elevator component are used as independent nodes to build a known elevator component failure action connection matrix.

[0060] Based on the transfer entropy, calculate the causal intensity entropy value between each pair of nodes in the known elevator component fault action connection matrix, and determine the causal chain between each pair of nodes.

[0061] Based on the known fault action linkage matrix of elevator components and the causal chain between each pair of nodes, a weighted directed graph of the causal network of known elevator component fault actions is constructed.

[0062] Based on Node2Vec random walk machine learning, an initial vector is assigned to each node in the weighted directed graph of the causal network of known elevator component failure actions, and the maximum random walk transition probability between the initial vectors of each node is calculated to generate a causal neighbor sequence of known elevator component failure actions.

[0063] Using the HDBSCAN clustering algorithm, we verify the prior knowledge between nodes in the causal neighbor sequence of known elevator component failure actions, determine the neighborhood radius and minimum density number to divide the nodes, generate the minimum spanning tree of the causal reachability metric of failure actions, and extract the causal association clusters of known elevator component failure actions.

[0064] Based on the known causal relationship clusters of elevator component failure actions, calculate the inter-cluster relationship with the largest causal flow between each pair of clusters as parent and child nodes, and build a fault tree diagram of the known elevator components.

[0065] When using this method, in conjunction with the steps mentioned above, it is important to note that the traditional method of constructing fault tree diagrams for elevator components relies too heavily on subjective experience, making it difficult to cover the hidden fault associations under complex underground working conditions. Furthermore, static rules are difficult to adapt to new fault modes during the dynamic operation of elevator equipment. The fine-grained and coarse-grained relationships of manually labeled fault nodes cannot quantify the causal strength, resulting in insufficient accuracy of the diagram in diagnosing complex faults. Consequently, the constructed fault tree diagrams for elevator components have a high false alarm rate in practical applications.

[0066] Therefore, by integrating known elevator component failure action data and utilizing natural language processing techniques, a weighted directed graph of the causal network of component failures in the target elevator is constructed. This graph is then combined with Node2Vec random walk and HDBSCAN clustering algorithms to uncover deep correlations in failure actions, ultimately generating a visualized fault tree graph. The beneficial effects include quantifying the strength of causal relationships using transfer entropy, and achieving automated hierarchical classification of failure modes through graph embedding and density clustering. This reduces the limitations of isolated data analysis in traditional fault diagnosis; it improves the efficiency of tracing complex faults by discovering hidden fault associations based on data-driven approaches; and it provides decision support for preventative maintenance by graphically representing fault propagation paths.

[0067] S2. Based on the historical operation log of the target elevator equipment, mark the historical operation process instructions of the target elevator equipment, analyze the matching between the historical operation process instructions and the assembled building, and evaluate the current status of each component of the target elevator equipment;

[0068] Step S2 includes the following:

[0069] Based on the historical operation logs of the target elevator equipment and the prefabricated building, extract the historical operation process instruction feature data of the target elevator equipment and the usage attribute feature data of the prefabricated building; the historical operation process instruction feature data includes: operation direction and operation speed, while the usage attributes of the prefabricated building include: building type, building functional zoning and building personnel flow.

[0070] Based on sliding window statistics, the runtime of the historical operation process instructions of the target elevator equipment is used as the observation window, and the unit time is used as the observation step. The historical operation process instruction feature data of the target elevator equipment is used as the observation object. The mean and variance of the historical operation process instructions within the sliding window are calculated to obtain the load status label of the historical operation process instructions of the target elevator equipment.

[0071] Based on principal component analysis, weights are assigned to the usage attribute characteristic data of prefabricated buildings, and the usage attribute characteristic data of prefabricated buildings are used to calculate the usage attribute characteristic vector index of prefabricated buildings.

[0072] The historical operation process instruction load status labels of the target elevator equipment and the feature vector index of the building's purpose attribute are normalized.

[0073] Based on DTW dynamic time warping, the cosine similarity between the historical operation process instruction load status label of the target elevator equipment and the feature vector index of the purpose attribute of the prefabricated building under a unit time is calculated to determine the matching degree between the historical operation process instruction of the target elevator equipment and the prefabricated building under a unit time.

[0074] Obtain historical standardized operating data of similar target elevator equipment in known prefabricated buildings;

[0075] Based on the historical standardized operation data of known prefabricated buildings and similar target elevator equipment, a GRU neural network is trained to establish a life prediction loss function for known prefabricated buildings and similar target elevator equipment. The historical operation process instruction feature data of the target elevator equipment and the usage attribute feature data of the prefabricated building are used as inputs, and the remaining health indicators of the target elevator equipment are used as outputs.

[0076] By using the historical operation process instructions of the target elevator equipment under a unit time and the matching degree of the assembled building, the remaining health indicators of the target elevator equipment are corrected to obtain the health status indicators of each component of the target elevator equipment.

[0077] When using it, refer to the steps outlined above.

[0078] As a further step, the matching degree between the usage attributes of prefabricated buildings and the DTW dynamic time warping time series of historical operating process command loads is quantified through sliding window statistics and entropy weight method. This is then combined with a GRU neural network to predict the remaining lifespan of target elevator equipment components, ultimately dynamically adjusting health indicators based on the matching degree. The technical principle lies in integrating multi-source time series data with the usage attribute parameters of prefabricated buildings, achieving cross-prefabricated building knowledge sharing through federated learning, and solving the problem of disconnect between operating conditions and building usage in traditional methods. Beneficial effects include: adjusting health indicators based on real-time matching degree, reducing misjudgments caused by sudden changes in building type; and achieving accurate prediction of remaining lifespan by combining load and building attribute characteristics using the GRU model, reducing unplanned downtime.

[0079] S3. Based on edge sensors, obtain the real-time operating data of each component of the target elevator equipment and perform similar fault screening with the known elevator component fault tree map to obtain the set of undetermined faults for each component of the real-time target elevator equipment.

[0080] Step S3 includes the following:

[0081] Based on the real-time operational data of each component of the target elevator equipment, multi-dimensional characteristic parameters of each component of the real-time target elevator equipment are marked; the multi-dimensional characteristic parameters include: vibration parameters, temperature parameters, and current and voltage parameters;

[0082] Based on the design service life and basic performance degradation coefficient of each component of the real-time target elevator equipment and the cumulative operating time of each component of the real-time target elevator equipment, the dynamic fault threshold of each component of the real-time target elevator equipment is estimated.

[0083] Using the analytic hierarchy process (AHP), weights are assigned to the multi-dimensional operational characteristic parameters of each component of the target elevator equipment. Based on the multi-dimensional operational characteristic parameters of each component of the target elevator equipment and the known fault tree diagram of the elevator components, the weighted fault similarity index of each component of the target elevator equipment is calculated.

[0084] The system selects fault actions whose weighted fault similarity index for each component of the target elevator equipment is greater than the dynamic fault threshold of each component of the target elevator equipment, and constructs a set of undetermined faults for each component of the target elevator equipment.

[0085] S4. Based on the current status of each component of the target elevator equipment and the set of undetermined faults of each component of the target elevator equipment, establish an elevator component-fault confidence function to generate the fault types of each component of the target elevator equipment.

[0086] Step S4 includes the following:

[0087] Verify the support value of the operating data of each component of the target elevator equipment for each undetermined fault in the set of undetermined faults of each component of the target elevator equipment. Using DS evidence theory, fuse the combined probabilities of the support values ​​of each undetermined fault in the set of undetermined faults of each component of the target elevator equipment to obtain the joint probability of the operating data of each component of the target elevator equipment for each undetermined fault of the target elevator equipment.

[0088] Based on the joint probability of undetermined faults of each component of the target elevator equipment and the fault tree diagram of known elevator components, the prior probability of faults of each component of the target elevator equipment is obtained.

[0089] Based on the operating data of each component of the target elevator equipment and the set of undetermined faults of each component of the target elevator equipment, a likelihood function is established to verify the likelihood probability of the operating data of each component of the target elevator equipment under the undetermined faults of each component.

[0090] Based on the prior probability of each component failure of the target elevator equipment and the likelihood probability of the operating data of each component of the target elevator equipment under the condition of undetermined failure of each component, the posterior confidence of each component failure of the target elevator equipment is calculated.

[0091] The health status indicators of each component of the target elevator equipment are normalized and used as the weights of the health status indicators of each component of the target elevator equipment.

[0092] Based on the weights of the health status indicators of each component of the target elevator equipment, the posterior confidence of the faults of each component of the target elevator equipment, and the set of undetermined faults of each component of the target elevator equipment, an elevator component status-fault confidence function is established to calculate the confidence of the fault type of each component of the target elevator equipment, as follows:

[0093] S i =α·P(F i |S)+β·w i ·γ·Severity(F i )

[0094] Among them, S i Let P(F) be the confidence level of the fault type of the i-th component of the target elevator equipment. i |S) represents the posterior confidence level of the failure of the i-th component given the operating data of each component of the target elevator equipment, w i Severity(F) represents the weight of the remaining health indicators for the i-th component of the target elevator equipment. i Let α be the risk index of the failure of the i-th component of the target elevator equipment, β be the health weight coefficient, and γ be the failure risk coefficient.

[0095] When using it, refer to the steps outlined above.

[0096] As a further step, this study utilizes DS evidence theory to fuse multi-sensor data and calculates the joint support probability of undetermined faults. Combining prior knowledge from fault tree mapping, Bayes' theorem is used to transform the real-time data likelihood probability into posterior confidence. Furthermore, equipment health status weights are introduced to construct an elevator state-fault confidence function for comprehensive scoring. The beneficial effects are: resolving multi-sensor data conflict issues through evidence theory, reducing false alarm rates; and prioritizing high-risk components through a health status weighting mechanism, improving maintenance efficiency and fault location, and reducing maintenance costs.

[0097] Reference Figure 2 As shown, a fault monitoring and diagnosis system for the entire lifecycle of elevator operation and maintenance is proposed, including:

[0098] Fault diagram module, elevator component estimation module, elevator fault screening module, and elevator fault location module;

[0099] The fault map module is used to collect known elevator component fault action data based on the Internet, analyze the correlation trends between elevator component fault action data, and establish elevator component fault tree map.

[0100] The elevator component estimation module is used to mark the historical operation process instructions of the target elevator equipment based on the historical operation log of the target elevator equipment, analyze the matching between the historical operation process instructions and the assembly building, and evaluate the current status of each component of the target elevator equipment.

[0101] The elevator fault screening module is electrically connected to the fault map module and the elevator component estimation module. The elevator fault screening module is used to obtain the real-time operating data of each component of the target elevator equipment based on edge sensors and perform similar fault screening with the known elevator component fault tree map to obtain the set of undetermined faults of each component of the real-time target elevator equipment.

[0102] The elevator fault location module is electrically connected to the elevator fault screening module. The elevator fault location module is used to establish an elevator component-fault confidence function based on the current status of each component of the target elevator equipment and the set of pending faults of each component of the target elevator equipment, and generate the fault types of each component of the target elevator equipment.

[0103] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for fault monitoring and diagnosis throughout the entire lifecycle of elevator operation and maintenance, characterized in that, include: S1. Based on the Internet, collect known elevator component failure action data, analyze the correlation trend between elevator component failure action data, and establish elevator component fault tree map. S2. Based on the historical operation log of the target elevator equipment, mark the historical operation process instructions of the target elevator equipment, analyze the matching between the historical operation process instructions and the assembled building, and evaluate the current status of each component of the target elevator equipment; S3. Based on edge sensors, obtain the real-time operating data of each component of the target elevator equipment and perform similar fault screening with the known elevator component fault tree map to obtain the set of undetermined faults for each component of the real-time target elevator equipment. S4. Based on the current status of each component of the target elevator equipment and the set of undetermined faults of each component of the target elevator equipment, establish an elevator component-fault confidence function to generate the fault types of each component of the target elevator equipment.

2. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 1, characterized in that, Step S1 includes the following: Based on the elevator equipment manufacturer's fault database and elevator equipment fault case library, obtain known elevator component fault action data; Based on known elevator component failure action data, NLP (Natural Language Processing) is used to extract the causal relationship between the failure phenomenon and the failure cause of the elevator component. The failure, failure phenomenon and failure cause of each elevator component are used as independent nodes to build a known elevator component failure action connection matrix. Based on the transfer entropy, calculate the causal intensity entropy value between each pair of nodes in the known elevator component fault action connection matrix, and determine the causal chain between each pair of nodes. Based on the known fault action linkage matrix of elevator components and the causal chain between each pair of nodes, a weighted directed graph of the causal network of known elevator component fault actions is constructed.

3. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 2, characterized in that, Step S1 further includes: Based on Node2Vec random walk machine learning, an initial vector is assigned to each node in the weighted directed graph of the causal network of known elevator component failure actions, and the maximum random walk transition probability between the initial vectors of each node is calculated to generate a causal neighbor sequence of known elevator component failure actions. Using the HDBSCAN clustering algorithm, we verify the prior knowledge between nodes in the causal neighbor sequence of known elevator component failure actions, determine the neighborhood radius and minimum density number to divide the nodes, generate the minimum spanning tree of the causal reachability metric of failure actions, and extract the causal association clusters of known elevator component failure actions. Based on the known causal relationship clusters of elevator component failure actions, calculate the inter-cluster relationship with the largest causal flow between each pair of clusters as parent and child nodes, and build a fault tree diagram of the known elevator components.

4. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 3, characterized in that, Step S2 includes the following: Based on the historical operation logs of the target elevator equipment and the prefabricated building, extract the historical operation process instruction feature data of the target elevator equipment and the usage attribute feature data of the prefabricated building; The historical operation process instruction feature data includes: operation direction and operation speed, while the usage attributes of the prefabricated building include: building type, building functional zoning, and building personnel flow. Based on sliding window statistics, the runtime of the historical operation process instructions of the target elevator equipment is used as the observation window, and the unit time is used as the observation step. The historical operation process instruction feature data of the target elevator equipment is used as the observation object. The mean and variance of the historical operation process instructions within the sliding window are calculated to obtain the load status label of the historical operation process instructions of the target elevator equipment. Based on principal component analysis, weights are assigned to the usage attribute feature data of prefabricated buildings, and the usage attribute feature vector index of prefabricated buildings is calculated.

5. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 4, characterized in that, Step S2 further includes: The historical operation process instruction load status labels of the target elevator equipment and the feature vector index of the building's purpose attribute are normalized. Based on DTW dynamic time warping, the cosine similarity between the historical operation process instruction load status label of the target elevator equipment and the feature vector index of the purpose attribute of the prefabricated building under a unit time is calculated to determine the matching degree between the historical operation process instruction of the target elevator equipment and the prefabricated building under a unit time. Obtain historical standardized operating data of similar target elevator equipment in known prefabricated buildings.

6. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 5, characterized in that, Step S2 further includes: Based on the historical standardized operation data of known prefabricated buildings and similar target elevator equipment, a GRU neural network is trained to establish a life prediction loss function for known prefabricated buildings and similar target elevator equipment. The historical operation process instruction feature data of the target elevator equipment and the usage attribute feature data of the prefabricated building are used as inputs, and the remaining health indicators of the target elevator equipment are used as outputs. By using the historical operating process instructions of the target elevator equipment per unit time and the assembly building matching degree, the remaining health indicators of the target elevator equipment are corrected to obtain the health status indicators of each component of the target elevator equipment.

7. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 6, characterized in that, Step S3 includes the following: Based on the real-time operational data of each component of the target elevator equipment, multi-dimensional characteristic parameters of each component of the real-time target elevator equipment are marked; the multi-dimensional characteristic parameters include: vibration parameters, temperature parameters, and current and voltage parameters; Based on the design service life and basic performance degradation coefficient of each component of the real-time target elevator equipment and the cumulative operating time of each component of the real-time target elevator equipment, the dynamic fault threshold of each component of the real-time target elevator equipment is estimated. Using the analytic hierarchy process (AHP), weights are assigned to the multi-dimensional operational characteristic parameters of each component of the target elevator equipment. Based on the multi-dimensional operational characteristic parameters of each component of the target elevator equipment and the known fault tree diagram of the elevator components, the weighted fault similarity index of each component of the target elevator equipment is calculated. The system selects fault actions whose weighted fault similarity index for each component of the target elevator equipment is greater than the dynamic fault threshold of each component of the target elevator equipment, and constructs a set of undetermined faults for each component of the target elevator equipment.

8. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 7, characterized in that, Step S4 includes the following: Verify the support value of the operating data of each component of the target elevator equipment for each undetermined fault in the set of undetermined faults of each component of the target elevator equipment. Using DS evidence theory, fuse the combined probabilities of the support values ​​of each undetermined fault in the set of undetermined faults of each component of the target elevator equipment to obtain the joint probability of the operating data of each component of the target elevator equipment for each undetermined fault of the target elevator equipment. Based on the joint probability of undetermined faults of each component of the target elevator equipment and the fault tree diagram of known elevator components, the prior probability of faults of each component of the target elevator equipment is obtained. Based on the operating data of each component of the target elevator equipment and the set of undetermined faults of each component of the target elevator equipment, a likelihood function is established to verify the likelihood probability of the operating data of each component of the target elevator equipment under the undetermined faults of each component. Based on the prior probability of each component failure of the target elevator equipment and the likelihood probability of the operating data of each component of the target elevator equipment under the condition of undetermined failure of each component, the posterior confidence of each component failure of the target elevator equipment is calculated.

9. The elevator operation and maintenance full life cycle fault monitoring and diagnosis method according to claim 8, characterized in that, Step S4 further includes: The health status indicators of each component of the target elevator equipment are normalized and used as the weights of the health status indicators of each component of the target elevator equipment. Based on the weights of the health status indicators of each component of the target elevator equipment, the posterior confidence of the faults of each component of the target elevator equipment, and the set of undetermined faults of each component of the target elevator equipment, an elevator component status-fault confidence function is established to calculate the confidence of the fault type of each component of the target elevator equipment, as follows: S i =α·P(F i |S)+β·w i ·γ·Severity(F i ) Among them, S i Let P(F) be the confidence level of the fault type of the i-th component of the target elevator equipment. i |S) represents the posterior confidence level of the failure of the i-th component given the operating data of each component of the target elevator equipment, w i Severity(F) represents the weight of the remaining health indicators for the i-th component of the target elevator equipment. i Let α be the risk index of the failure of the i-th component of the target elevator equipment, β be the health weight coefficient, and γ be the failure risk coefficient.

10. A fault monitoring and diagnosis system for the entire lifecycle of elevator operation and maintenance, characterized in that, To implement the elevator operation and maintenance full life cycle fault monitoring and diagnosis method as described in any one of claims 1-9, the method includes: Fault diagram module, elevator component estimation module, elevator fault screening module, and elevator fault location module; The fault map module is used to collect known elevator component fault action data based on the Internet, analyze the correlation trends between elevator component fault action data, and establish elevator component fault tree map. The elevator component estimation module is used to mark the historical operation process instructions of the target elevator equipment based on the historical operation log of the target elevator equipment, analyze the matching between the historical operation process instructions and the assembly building, and evaluate the current status of each component of the target elevator equipment. The elevator fault screening module is electrically connected to the fault map module and the elevator component estimation module. The elevator fault screening module is used to obtain the real-time operating data of each component of the target elevator equipment based on edge sensors and perform similar fault screening with the known elevator component fault tree map to obtain the set of undetermined faults of each component of the real-time target elevator equipment. The elevator fault location module is electrically connected to the elevator fault screening module. The elevator fault location module is used to establish an elevator component-fault confidence function based on the current status of each component of the target elevator equipment and the set of pending faults of each component of the target elevator equipment, and generate the fault types of each component of the target elevator equipment.

Images