An elevator fault occurrence probability calculation method and device based on multi-source data

Abstract

This invention relates to a method and apparatus for calculating the probability of elevator malfunctions based on multi-source data. The method generates elevator maintenance data using a trained XGBoost model; it trains a long short-term memory neural network model using automatically collected elevator operation data, elevator trigger data, offline statistical data, and the XGBoost-generated elevator maintenance data; then, it inputs the automatically collected elevator operation data, elevator trigger data, and offline statistical data into the trained long short-term memory neural network model to automatically obtain the elevator maintenance score HI, and then calculates the probability of elevator malfunctions. The entire process requires no manual intervention, and the calculation results are fast and accurate.

Description

Technical Field

[0001] This invention relates to the calculation of elevator failure probability, and in particular to a method and apparatus for calculating elevator failure probability based on multi-source data. Background Technology

[0002] Elevators are the most common mode of transportation in high-rise buildings, carrying tens of thousands of people daily. Due to the large number and frequency of elevator use, various types of malfunctions are inevitable. These include entrapment, overshooting, and emergency stops, all of which can pose safety risks to passengers. Door operator malfunctions and unresponsive buttons negatively impact passenger experience and elevator efficiency. Currently, elevator malfunctions are detected through passenger use or routine maintenance, a process that incurs significant manpower and time costs. With the development of big data and the Internet of Things, elevator operational and trigger data can be acquired in real-time. Furthermore, the rapid iteration and upgrading of algorithms in deep learning and machine learning allow for the analysis of the relationships between elevator components and their performance status. This enables the establishment of models linking elevator parameters and malfunctions, allowing for the prediction of elevator component failures and providing information for maintenance decisions.

[0003] In the current field of elevator fault prediction, most existing technologies tend to use direct parameters of the elevator's remaining lifespan or specific fault types (such as the horizontal vibration amplitude of the traction machine) as prediction indicators. While this approach is direct, it lacks flexibility and is difficult to fully cover the complex and varied elevator types and their diverse fault modes.

[0004] For example, in the patent "An Elevator Fault Prediction System and Method (CN117864892A)," the system mainly achieves full-process detection of elevator fault repair and handling by constructing a data acquisition module, a data preprocessing module, a fault prediction module, and an emergency handling module. The data acquisition module primarily acquires elevator operation data and elevator maintenance data through IoT sensing devices. Operation data mainly includes sensor data such as up / down signal data, door zone signal data, and car switch signal data; maintenance data includes fault data such as insufficient contactor voltage and SWD electronic board faults. The data preprocessing module analyzes the two types of data to derive the correlation data of elevator operation processes that lead to elevator maintenance data. The fault prediction module constructs an RBF neural network model based on the operation and maintenance data, employing the nearest neighbor algorithm to obtain the weights from the hidden layer to the output layer nodes. Using elevator operation data as input and elevator maintenance data as output, the model results are compared with data from similar elevators to determine whether the elevator fault is directly related to the operation data. When the predicted value is within the range of similar elevator maintenance data, an algorithm such as ARIMA is used to predict the elevator's remaining lifespan. The remaining lifespan is updated to elevator maintenance time, completing a closed loop from elevator failure prediction to on-site elevator maintenance. Summary of the Invention

[0005] This invention generates elevator maintenance data using a trained XGBoost model; it trains a long short-term memory neural network model using automatically collected elevator operation data, elevator trigger data, offline statistical data, and the elevator maintenance data generated by the XGBoost model; then it inputs the automatically collected elevator operation data, elevator trigger data, and offline statistical data into the trained long short-term memory neural network model to automatically obtain the elevator maintenance score HI, and then calculates the probability of elevator failure. The entire process requires no manual intervention, and the calculation results are fast and accurate.

[0006] This invention provides the following technical solution: a method for calculating the probability of elevator malfunctions based on multi-source data, comprising the following steps:

[0007] S1. Collect the first elevator operation data, the first elevator trigger data, and the first offline statistical data;

[0008] S2. Process the abnormal data in the first elevator operation data, the first elevator trigger data and the first offline statistical data to obtain the second elevator operation data, the second elevator trigger data and the second offline statistical data.

[0009] S3. Input the second elevator operation data, the second elevator trigger data, and the second offline statistical data into the long short-term memory neural network model to obtain the first elevator maintenance data. The first elevator maintenance data includes the following types of data: elevator number and elevator maintenance score (HI).

[0010] S4. Calculate the probability of failure f(HI) using the following formula:

[0011] f(HI) = x × e -HI / y ;

[0012] Where x is the maximum probability of failure, y is the degree to which the failure probability decreases, and e is the natural constant. The values ​​of x and y can be set according to the actual situation; generally, x is 100 and y is 30.

[0013] The first elevator operation data, the first elevator trigger data, and the first offline statistical data are all related, meaning they are collected based on the same batch of elevators.

[0014] Furthermore, the Long Short-Term Memory (LSTM) neural network model is trained through the following steps:

[0015] (1) Collect alarm information from the first elevator and maintenance data from the second elevator, and divide them into a first training set and a first test set;

[0016] Generally, the ratio of the first training set to the first test set is 7:3, but it can also be divided according to needs.

[0017] The alarm information of the first elevator and the maintenance data of the second elevator are related, that is, they are collected based on the same batch of elevators.

[0018] (2) The initial XGBoost model is trained using the first training set to obtain the trained XGBoost model; the first elevator alarm information in the first training set is used as the input of the XGBoost model, and the second elevator maintenance data in the first training set is used as the output of the XGBoost model.

[0019] (3) Test XGBoost using the first test set. If the accuracy reaches or exceeds the first threshold, the training is complete. If the accuracy is less than the first threshold, the first training set and the first test set are re-divided and the training is performed again.

[0020] Generally, the first threshold is 80%, but it can be set as needed.

[0021] (4) Collect the third elevator operation data, the third elevator trigger data, the third offline statistical data, and the second elevator alarm information, and divide them into the second training set and the second test set;

[0022] The collected data on the operation of the third elevator, the trigger data of the third elevator, the offline statistics data of the third elevator, and the alarm information of the second elevator are all related, that is, they are collected based on the same batch of elevators.

[0023] Generally, the ratio of the second training set to the second test set is 7:3, but it can also be divided according to needs.

[0024] (5) Input the second elevator alarm information from the second training set into the XGBoost model to obtain the third elevator maintenance data;

[0025] (6) The initial long short-term memory neural network model is trained using the third elevator operation data, third elevator trigger data, third offline statistical data and third elevator maintenance data in the second training set, and the trained long short-term memory neural network model is obtained; the third elevator operation data, third elevator trigger data and third offline statistical data in the second training set are used as input to the long short-term memory neural network model, and the third elevator maintenance data are used as output to the long short-term memory neural network model.

[0026] (7) The long short-term memory neural network model is tested using the second test set. If the accuracy reaches the second threshold or above, the training is completed. If the accuracy is less than the second threshold, the second training set and the second test set are re-divided and the training is repeated.

[0027] Generally, the second threshold is not 80%, but it can be set as needed.

[0028] Furthermore, the first elevator operation data, the second elevator operation data, and the third elevator operation data all include the following types of data: elevator operating speed (divided into x, y, and z axes), operating acceleration (divided into x, y, and z axes), temperature (divided into inside the car, top, and bottom), and humidity (divided into inside the car, top, and bottom).

[0029] The trigger data for the first elevator, the second elevator, and the third elevator all include the following types of data: elevator number, car door open, car door closed, current floor, running contactor engaged, and sealing contactor released.

[0030] The first, second, and third offline statistical data all include the following data types: elevator number, years of use, days of use, and total historical running distance;

[0031] The second elevator maintenance data includes the following types of data: elevator number and elevator maintenance score HI. The elevator maintenance score HI is the professional assessment of the current condition of the elevator by the operation and maintenance personnel (according to Tables 1 to 4).

[0032] Table 1. Elevator Maintenance Dimensions

[0033]

[0034]

[0035]

[0036] Table 2 Probability Assessment Table

[0037]

[0038] Table 3 Risk Assessment Table

[0039]

[0040] Table 4 Maintenance and Repair Score Evaluation Table

[0041]

[0042] Both the first and second elevator alarm information include the following types of data: elevator alarm information includes elevator number, elevator alarm start time, elevator alarm end time, elevator status, elevator alarm type code, and number of alarm occurrences;

[0043] The elevator alarm type code can be associated with the elevator alarm type code table to obtain the elevator alarm type name, fault type, and alarm type. The number of alarm occurrences is divided by hour. If the start and end times of an alarm are within the current hourly time interval, the number of alarm occurrences in that hourly time interval is increased by one. If the start and end times of an alarm span multiple hourly time intervals, the number of alarm occurrences in each of those multiple hourly time intervals is increased by one.

[0044] The number of alarm occurrences is divided into hours. If the start and end times of an alarm are within the current one-hour time interval, the number of alarm occurrences within that one-hour time interval is increased by one. If the start and end times of an alarm span multiple hourly time intervals, the number of alarm occurrences within each of those multiple hourly time intervals is increased by one.

[0045] Furthermore, the elevator maintenance score (HI) includes individual maintenance scores for six dimensions: car, lobby, car top, pit bottom, machine room, and landing doors. The individual maintenance score for each dimension is calculated using the following formula:

[0046] Single-dimensional maintenance score = 100 × (v1 + v2 + ... + vn) / 2n;

[0047] Where n is the total number of evaluation items in a single dimension, and v1, v2, ..., vn are the scores of the 1st to nth evaluation items in a single dimension.

[0048] An elevator fault occurrence probability calculation device based on multi-source data according to claim 1, comprising at least:

[0049] The first acquisition module is used to collect the first elevator operation data, the first elevator trigger data, and the first offline statistical data.

[0050] The first processing module is used to process abnormal data in the first elevator operation data, the first elevator trigger data and the first offline statistical data to obtain the second elevator operation data, the second elevator trigger data and the second offline statistical data.

[0051] The first analysis module is used to input the second elevator operation data, the second elevator trigger data, and the second offline statistical data into the long short-term memory neural network model to obtain the first elevator maintenance data. The first elevator maintenance data includes the following types of data: elevator number and elevator maintenance score (HI).

[0052] The first calculation module calculates the probability of failure f(HI) using the following formula:

[0053] f(HI) = x × e -HI / y ;

[0054] Where x is the maximum probability of failure, y is the degree of decrease in the failure probability, and e is the natural constant.

[0055] An electronic device, comprising:

[0056] One or more processors;

[0057] Memory, used to store one or more programs;

[0058] When the one or more programs are executed by the one or more processors, the one or more processors perform the methods described above.

[0059] A computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of the method described above.

[0060] The beneficial effects of this invention are as follows:

[0061] By using deep learning models, the health status of elevators can be predicted, which can be converted into the probability of elevator failure. Furthermore, various elevator health indices can be constructed for different components or failure types to predict the probability of various elevator failures. Attached Figure Description

[0062] Figure 1 This is a flowchart of the method of the present invention;

[0063] Figure 2This is a schematic diagram of the device module of the present invention;

[0064] The module includes a first acquisition module 1, a first processing module 2, a first analysis module 3, and a first calculation module 4. Detailed Implementation

[0065] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0066] This embodiment, based on real-time operational data, elevator trigger data, and offline statistical data from the past 15 days, predicts the probability of various faults occurring hourly for the next day, with an hourly granularity. The designed operational cycle is 6 hours. It achieves automatic learning of the deep correlation between elevator alarm information and maintenance records, reducing reliance on expert experience and resulting in accurate prediction of various elevator faults. Specifically:

[0067] I. Training the XGBoost model:

[0068] (1) The first elevator alarm information is composed of data such as the number of times 17 types of elevator alarms occurred in 20 elevators in 24 hours every 15 days, and the second elevator maintenance data is composed of data such as the maintenance HI of 20 elevators in 24 hours every 15 days. They are divided into the first training set and the first test set in a 7:3 ratio.

[0069] (2) The initial XGBoost model is trained using the first training set to obtain the trained XGBoost model; the first elevator alarm information in the first training set is used as the input of the XGBoost model, and the second elevator maintenance data in the first training set is used as the output of the XGBoost model.

[0070] The XGBoost model obtains its hyperparameters using a conventional grid search method. These hyperparameters include the learning rate, maximum depth, subsample ratio, feature sampling ratio, and minimum child node weight. Early stopping is used to prevent overfitting. The efficient and scalable XGBoost gradient boosting framework is employed for in-depth mining and analysis of massive elevator alarm and maintenance data. XGBoost not only effectively handles high-dimensional sparse data but also prevents overfitting through its built-in regularization and pruning strategies, ensuring the model's generalization ability. XGBoost's automatic feature selection and weight allocation eliminate the need for manually defining complex metric rules.

[0071] (3) Test XGBoost using the first test set. If the accuracy reaches 80% or above, the training is complete. If the accuracy is less than 80%, the first training set and the first test set are re-divided and the training is repeated.

[0072] (4) The third elevator operation data is composed of 12 operation data feature fields from 200 elevators extracted from records every 24 hours for 15 days; the third elevator trigger data is composed of 5 elevator trigger data feature fields from 200 elevators every 24 hours for 15 days; the third offline statistical data is composed of 3 elevator offline statistical data feature fields from 200 elevators every 24 hours for 15 days; and the second elevator alarm information is composed of 17 types of alarm statistical fields from 200 elevators every 24 hours for 15 days. The data is divided into the second training set and the second test set in a 7:3 ratio.

[0073] II. Training the Long Short-Term Memory Neural Network Model:

[0074] (5) Input the second elevator alarm information from the second training set into the XGBoost model to obtain the third elevator maintenance data;

[0075] (6) The initial long short-term memory neural network model is trained using the third elevator operation data, third elevator trigger data, third offline statistical data and third elevator maintenance data in the second training set, and the trained long short-term memory neural network model is obtained; the third elevator operation data, third elevator trigger data and third offline statistical data in the second training set are used as input to the long short-term memory neural network model, and the third elevator maintenance data are used as output to the long short-term memory neural network model.

[0076] (7) The long short-term memory neural network model is tested using the second test set. If the accuracy reaches 80% or above, the training is complete. If the accuracy is less than 80%, the second training set and the second test set are re-divided and the training is repeated.

[0077] Using the HI health index calculated through XGBoost as the training label for the bidirectional LSTM neural network model further improves the accuracy and efficiency of fault prediction. Deep learning models can capture nonlinear relationships in data, and combined with the comprehensive HI index, they can more accurately predict the types and timing of future elevator faults.

[0078] III. Calculating the probability of elevator malfunction:

[0079] A method for calculating the probability of elevator malfunctions based on multi-source data includes the following steps:

[0080] S1. The first elevator operation data consists of 12 operation data feature fields from 500 elevators recorded every 24 hours for 15 days; the first elevator trigger data consists of 5 elevator trigger data feature fields from 500 elevators recorded every 24 hours for 15 days; and the first offline statistical data consists of 3 elevator offline statistical data feature fields from 500 elevators recorded every 24 hours for 15 days.

[0081] S2. Process the abnormal data in the first elevator operation data, the first elevator trigger data and the first offline statistical data to obtain the second elevator operation data, the second elevator trigger data and the second offline statistical data. The dimensions of the second elevator operation data, the second elevator trigger data and the second offline statistical data after abnormal data processing should be the same as the corresponding dataset a.

[0082] In this embodiment, the abnormal data is handled using a box plot statistical method.

[0083] The specific method is to find the first quartile Q1, median Q2, and third quartile Q3 in the data to be processed, calculate the interquartile range IQR = Q3 - Q1, calculate the lower threshold Q1 - 1.5 * IQR and the upper threshold Q3 + 1.5 * IQR, and identify and delete data in the data to be processed that exceeds the upper threshold or is below the lower threshold as abnormal data.

[0084] S3. Input the second elevator operation data, the second elevator trigger data, and the second offline statistical data into the long short-term memory neural network model to obtain the first elevator maintenance data. The first elevator maintenance data includes elevator maintenance data HI.

[0085] S4. Calculate the probability of failure f(HI) using the following formula:

[0086] f(HI) = x × e -HI / y ;

[0087] Where x is the maximum probability of failure, y is the degree of decrease in the probability of failure, x is 100, y is 30, and e is the natural constant.

[0088] The method for verifying this application is as follows:

[0089] Failures with a probability of 50% or greater are considered high-probability failures requiring maintenance, while failures with a probability of less than 50% are considered low-probability failures requiring no maintenance for the time being. The predicted failure probability f(HI) is converted into a binary result. By comparing this result with the actual failures occurring within the predicted timeframe, the confusion matrix is ​​calculated, yielding TP (both predicted and actual failures), TN (both predicted and actual failures), FP (predicted failures, actual failures), and FN (predicted failures, actual failures). This leads to the calculation of the recall rate TPR = (TP / (TP+FN)) and the false positive rate FPR = (FP / (FP+TN)). The false positive rates FPR are then sorted in ascending order and substituted into the formula to calculate the model's AUC value. Where n is the predicted number of 500 elevators over 24 hours, with an hourly granularity, i.e., n = 12000. When the AUC value is greater than 0.7, it indicates that the failure probability calculation of this application is accurate.

[0090] An elevator fault occurrence probability calculation device based on multi-source data according to the above method, comprising at least:

[0091] The first acquisition module is used to collect the first elevator operation data, the first elevator trigger data, and the first offline statistical data.

[0092] The first processing module is used to process abnormal data in the first elevator operation data, the first elevator trigger data and the first offline statistical data to obtain the second elevator operation data, the second elevator trigger data and the second offline statistical data.

[0093] The first analysis module is used to input the second elevator operation data, the second elevator trigger data, and the second offline statistical data into the long short-term memory neural network model to obtain the first elevator maintenance data, which includes elevator maintenance data (HI).

[0094] The first calculation module calculates the probability of failure f(HI) using the following formula:

[0095] f(HI) = x × e - HI / y;

[0096] Where x is the maximum probability of failure, y is the degree of decrease in the failure probability, and e is the natural constant.

[0097] An electronic device, comprising:

[0098] One or more processors;

[0099] Memory, used to store one or more programs;

[0100] When the one or more programs are executed by the one or more processors, the one or more processors perform the methods described above.

[0101] In the embodiments provided in this application, it should be understood that the disclosed methods and systems can also be implemented in other ways. The method and system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0102] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0103] On the other hand, a computer-readable storage medium stores computer instructions thereon, which, when executed by a processor, implement the steps of the methods described above. When the computer program is executed by a processor, it implements the methods described in any of the first aspects above. If the functionality is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0104] In this application embodiment, on the one hand, by installing multiple types of sensors in the elevator, the fusion of multi-sensor information helps to reduce environmental interference that may affect a single sensor, thereby improving data reliability; on the other hand, by using the XGBoost model and the Long Short-Term Memory Neural Network model to fuse the feature information of the elevator operating status, the accuracy of the elevator fault probability calculation can be improved, thereby reducing false alarms and missed alarms, ensuring the efficient operation of the elevator fault probability calculation, significantly improving elevator safety, and reducing the probability of accidents.

[0105] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for calculating the probability of elevator failure occurrence based on multi-source data, characterized in that, The method comprises the following steps: S1, collecting first elevator operation data, first elevator trigger data and first offline statistical data; S2, processing abnormal data in the first elevator operation data, the first elevator trigger data and the first offline statistical data to obtain second elevator operation data, second elevator trigger data and second offline statistical data; S3, inputting the second elevator operation data, the second elevator trigger data and the second offline statistical data into a long short-term memory neural network model to obtain first elevator maintenance data, the first elevator maintenance data comprising the following types of data: elevator number and elevator maintenance grade HI; S4, calculating the failure occurrence probability f(HI) by the following formula: f(HI) = x x e -HI / y ; Wherein, x is the maximum probability of failure, y is the degree of decline of the failure probability, and e is a natural constant; The long short-term memory neural network model is trained by the following steps: (1) Collecting first elevator alarm information and second elevator maintenance data, and dividing them into a first training set and a first test set; (2) Training an initial XGBoost model using the first training set to obtain a trained XGBoost model; inputting the first elevator alarm information in the first training set into the XGBoost model, and inputting the second elevator maintenance data in the first training set into the XGBoost model; (3) Testing the XGBoost using the first test set, if the accuracy rate reaches or is higher than a first threshold value, the training is completed, if the accuracy rate is lower than the first threshold value, the first training set and the first test set are re-divided and the training is re-performed; (4) Collecting third elevator operation data, third elevator trigger data, third offline statistical data and second elevator alarm information, and dividing them into a second training set and a second test set; (5) Inputting the second elevator alarm information in the second training set into the XGBoost model to obtain third elevator maintenance data; (6) Training an initial long short-term memory neural network model using the third elevator operation data, the third elevator trigger data, the third offline statistical data and the third elevator maintenance data in the second training set to obtain a trained long short-term memory neural network model; inputting the third elevator operation data, the third elevator trigger data and the third offline statistical data in the second training set into the long short-term memory neural network model, and inputting the third elevator maintenance data into the long short-term memory neural network model; (7) Testing the long short-term memory neural network model using the second test set, if the accuracy rate reaches or is higher than a second threshold value, the training is completed, if the accuracy rate is lower than the second threshold value, the second training set and the second test set are re-divided and the training is re-performed; The first elevator operation data, the second elevator operation data and the third elevator operation data all comprise the following types of data: elevator running speed, running acceleration, temperature and humidity; The first elevator trigger data, the second elevator trigger data and the third elevator trigger data all comprise the following types of data: elevator number, car door opening, car door closing, current floor, running contactor attraction and star contactor release. The first offline statistical data, the second offline statistical data, and the third offline statistical data each include the following types of data: elevator number, service years, service days, and historical total running distance; The second elevator maintenance data and the second elevator maintenance data each include the following types of data: elevator number and elevator maintenance score HI; The first elevator alarm information and the second elevator alarm information each include the following types of data: elevator number, elevator alarm start time, elevator alarm end time, elevator state, elevator alarm type code, and alarm occurrence number; The elevator maintenance score HI includes single-dimension maintenance scores of six dimensions of car, hall, car roof, pit bottom, machine room, and landing door, and each single-dimension maintenance score is calculated by the following formula: Single-dimension maintenance score = 100 x (v1 + v2 + … + vn) / 2n; Wherein, n is the total number of evaluation items in the single dimension, and v1, v2, …, vn are scores of the first to nth evaluation items in the single dimension.

2. The method of claim 1, wherein, The abnormal data is processed by a box plot statistical method.

3. A multi-source data-based elevator failure occurrence probability calculation device according to the method of claim 1, characterized by, At least comprising: A first collection module configured to collect first elevator running data, first elevator trigger data, and first offline statistical data; A first processing module configured to process abnormal data in the first elevator running data, the first elevator trigger data, and the first offline statistical data to obtain second elevator running data, second elevator trigger data, and second offline statistical data; A first analysis module configured to input the second elevator running data, the second elevator trigger data, and the second offline statistical data into a long short-term memory neural network model to obtain first elevator maintenance data, the first elevator maintenance data including the following types of data: elevator number and elevator maintenance score HI; A first calculation module configured to calculate a failure occurrence probability f(HI) by the following formula: f(HI) = x x e -HI / y ; Wherein, x is the maximum probability of failure, y is the degree of decline of the failure probability, and e is a natural constant.

4. An electronic device, comprising: Comprising: One or more processors; Memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1-2.

5. A computer readable storage medium having stored thereon computer instructions, wherein, The instructions are executed by the processor to implement the steps of the method of any one of claims 1-2.

Images