Centralized apartment management system and management method thereof

Abstract

The embodiment of the invention provides a centralized apartment management system and a management method thereof, and belongs to the technical field of apartment management. The system comprises a customer service module, a data analysis module and an overhaul and maintenance module, and the customer service module is used for obtaining an overhaul request of a target user and analyzing the overhaul request to obtain target equipment and a target identifier of the target equipment; the data analysis module is used for obtaining initial operation data of the target equipment from the database according to the target identifier, and carrying out anomaly identification on the initial operation data to obtain target operation data of the target equipment; the equipment maintenance module is used for performing abnormal feature extraction on the target operation data to obtain a target feature vector, and performing abnormal classification according to the target feature vector to obtain a target fault type of the target equipment; determining an exception handling strategy of the maintenance request according to the target fault type; and sending the exception handling strategy and the maintenance request to a maintainer, so that the maintainer performs maintenance processing on the maintenance request according to the exception handling strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of apartment management, and in particular to a centralized apartment management system and a management method thereof. Background Art

[0002] As the scale of modern apartment management continues to expand, traditional manual management methods have become difficult to cope with the increasingly complex operation and maintenance needs of apartments. In particular, in the equipment inspection and maintenance of apartments, the traditional model mainly relies on manual inspections and user repair reports, which often results in untimely responses, inaccurate fault location, and incomplete problem handling. In addition, with the increasing number of smart devices used in apartments (such as smart door locks, water and electricity meters, etc.), the detection and maintenance needs of equipment failures have gradually become more complicated, placing higher requirements on the response speed and problem handling capabilities of maintenance personnel.

[0003] Therefore, there is an urgent need for a centralized apartment management system to improve the efficiency and accuracy of apartment equipment maintenance and user demand response. Summary of the invention

[0004] The main purpose of the embodiments of the present invention is to provide a centralized apartment management system and a management method thereof, aiming to solve the problems of low efficiency and insufficient accuracy in the process of apartment equipment maintenance and user demand response in the related technology.

[0005] In a first aspect, an embodiment of the present invention provides a centralized apartment management system, which includes a customer service module, a data analysis module, and a repair and maintenance module, wherein:

[0006] The customer service module is used to obtain a maintenance request corresponding to a target user, and parse the maintenance request to obtain a target device and a target identifier corresponding to the target device;

[0007] The data analysis module is used to obtain the initial operation data corresponding to the target device from the database according to the target identifier, and perform abnormality identification on the initial operation data to obtain the target operation data corresponding to the target device;

[0008] The equipment maintenance module is used to extract abnormal features from the target operating data to obtain a corresponding target feature vector, and to perform abnormal classification based on the target feature vector to obtain a target fault type corresponding to the target equipment; determine an abnormality handling strategy corresponding to the maintenance request based on the target fault type; and send the abnormality handling strategy and the maintenance request to maintenance personnel, so that the maintenance personnel can perform maintenance processing on the maintenance request based on the abnormality handling strategy.

[0009] In a second aspect, an embodiment of the present invention provides a centralized apartment management method, including:

[0010] Obtaining a maintenance request corresponding to a target user, and parsing the maintenance request to obtain a target device and a target identifier corresponding to the target device;

[0011] Obtaining initial operation data corresponding to the target device from a database according to the target identifier, and performing abnormality identification on the initial operation data to obtain target operation data corresponding to the target device;

[0012] Perform abnormal feature extraction on the target operation data to obtain a corresponding target feature vector, and perform abnormal classification based on the target feature vector to obtain a target fault type corresponding to the target equipment; determine an abnormal handling strategy corresponding to the maintenance request based on the target fault type; send the abnormal handling strategy and the maintenance request to maintenance personnel, so that the maintenance personnel perform maintenance processing on the maintenance request according to the abnormal handling strategy.

[0013] In a third aspect, an embodiment of the present invention further provides a terminal device, comprising a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one of the centralized apartment management methods provided in the specification of the present invention are implemented.

[0014] In a fourth aspect, an embodiment of the present invention further provides a storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any centralized apartment management method provided in the specification of the present invention.

[0015] The embodiment of the present invention provides a centralized apartment management system and a management method thereof. The centralized apartment management system includes a customer service module, a data analysis module, and a maintenance and repair module. After obtaining the maintenance request corresponding to the target user, the customer service module in the system parses the maintenance request to obtain the target device and the target identifier corresponding to the target device, and then uses the data analysis module to obtain the initial operation data corresponding to the target device from the database according to the target identifier, and then performs abnormal identification on the initial operation data to obtain the target operation data corresponding to the target device, so that the equipment maintenance module extracts abnormal features from the target operation data to obtain the corresponding target feature vector, and performs abnormal classification according to the target feature vector to obtain the target fault type corresponding to the target device; determines the abnormal handling strategy corresponding to the maintenance request according to the target fault type; sends the abnormal handling strategy and the maintenance request to the maintenance personnel, so that the maintenance personnel performs maintenance processing on the maintenance request according to the abnormal handling strategy. The system uses the data analysis module to perform abnormal identification on the initial operation data, generates the target operation data, and further improves the accuracy of fault judgment. The equipment maintenance module extracts abnormal features from the target operation data to form a target feature vector, which helps to quickly identify and classify the abnormal state of the equipment. According to the target feature vector, the abnormality is classified, the specific fault type of the target equipment is determined, and the intelligent diagnosis of the fault is realized. According to the target fault type, the optimal abnormality handling strategy can be generated to reduce the deviation and delay of human judgment. The abnormality handling strategy and maintenance request are sent directly to the maintenance personnel to ensure that they can perform maintenance quickly and accurately, and improve service efficiency. The efficiency and accuracy of apartment equipment maintenance are significantly improved, the user demand response process is optimized, the fault handling time is reduced, and customer satisfaction is improved. The problem of low efficiency and insufficient accuracy in the process of apartment equipment maintenance and user demand response in related technologies is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A schematic diagram of the module structure of a centralized apartment management system provided by an embodiment of the present invention;

[0018] Figure 2 A flowchart of a centralized apartment management method provided by an embodiment of the present invention;

[0019] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0022] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0023] The embodiment of the present invention provides a centralized apartment management system and a management method thereof. The centralized apartment management method can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0024] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0025] Please refer to Figure 1 , Figure 1 A schematic diagram of the module structure of a centralized apartment management system provided by an embodiment of the present invention.

[0026] like Figure 1As shown, the centralized apartment management system 200 includes a customer service module 201, a data analysis module 202, and a maintenance module 203. The customer service module 201 is used to obtain a maintenance request corresponding to a target user, and parse the maintenance request to obtain a target device and a target identifier corresponding to the target device; the data analysis module 202 is used to obtain the initial operation data corresponding to the target device from the database according to the target identifier, and perform abnormality identification on the initial operation data to obtain the target operation data corresponding to the target device; the equipment maintenance module 203 is used to extract abnormal features from the target operation data to obtain a corresponding target feature vector, and perform abnormal classification according to the target feature vector to obtain a target fault type corresponding to the target device; determine the abnormality handling strategy corresponding to the maintenance request according to the target fault type; send the abnormality handling strategy and the maintenance request to the maintenance personnel, so that the maintenance personnel perform maintenance processing on the maintenance request according to the abnormality handling strategy.

[0027] Exemplarily, the centralized apartment management system 200 includes a customer service module 201 , a data analysis module 202 , and a repair and maintenance module 203 .

[0028] For example, the centralized apartment management system 200 integrates multiple functional modules such as the customer service module 201, the data analysis module 202, and the inspection and maintenance module 203, and centralizes all management functions of the apartment on a single platform for unified management. This centralized management platform enables the centralized apartment management system 200 to coordinate resources, manage equipment, and process user requests across multiple apartment projects without the need to deploy a separate system for each apartment.

[0029] Exemplarily, the target user has a client installed in the mobile terminal, and sends a maintenance request to the customer service module 201 through the client, so that the customer service module 201 obtains the maintenance request corresponding to the target user, and then the customer service module 201 parses the maintenance request to obtain the target device that needs to be inspected or repaired, and then obtains the user identifier corresponding to the target user from the maintenance request, and then queries the database according to the user identifier and the device name corresponding to the target device to obtain the target identifier corresponding to the target device.

[0030] For example, all maintenance requests of target users are parsed through a unified customer service module, and the centralized apartment management system 200 automatically retrieves corresponding equipment data and fault information for processing according to these centrally managed requests. This approach avoids the decentralized management approach of each apartment independently processing user requests, and improves the overall responsiveness of the system.

[0031] Exemplarily, the centralized apartment management system 200 collects operation data corresponding to various devices in the apartment through sensors and processes them in a centralized data platform. By centralizing the operation data of all devices in a database, the centralized apartment management system 200 can perform comprehensive analysis and anomaly detection. Centralized data processing can not only improve data processing efficiency, but also timely query the initial operation data corresponding to the target device.

[0032] Exemplarily, the data analysis module 202 performs data query from the database according to the target identifier to obtain the initial operation data corresponding to the target device.

[0033] Exemplarily, the data analysis module 202 cleans and preprocesses the initial operation data to remove noise and invalid data to ensure the integrity and consistency of the data. Thus, the preprocessed data is analyzed using statistical analysis methods or machine learning models to identify abnormal data in the initial operation data, that is, to obtain target operation data.

[0034] For example, you can choose commonly used anomaly detection algorithms, such as statistical methods (such as Z-score, 3σ principle), time series analysis (such as ARIMA model), cluster analysis (such as K-means), or deep learning methods (such as autoencoder, LSTM), etc., and select a suitable model according to the characteristics of the target device. The preprocessed data is then input into the established model, and the abnormal data points are identified, thereby obtaining the target operation data corresponding to the target device.

[0035] Exemplarily, the equipment maintenance module 203 selects features that are closely related to the performance of the target equipment from the initial operation data. These features should be able to reflect the difference between normal and abnormal states. For example, statistical properties such as the mean, variance, extreme value, and rate of change of the data can be focused on. In-depth analysis of the selected features may require conversion of certain features (such as normalization and standardization) to adapt to subsequent classification algorithms. Methods such as principal component analysis (PCA) can be used to reduce data dimensions while retaining most of the information. The extracted abnormal features are integrated into a target feature vector.

[0036] Exemplarily, the equipment maintenance module 203 selects a suitable machine learning or statistical classification algorithm, such as support vector machine (SVM), decision tree, random forest, neural network, etc., according to the characteristics of the feature vector. The classification model is trained using a labeled feature vector set to establish a mapping relationship between features and fault types. Thus, the target feature vector is input into the trained model for abnormal classification, and the corresponding target fault type is output.

[0037] Exemplarily, the equipment maintenance module 203 searches the policy mapping table for the target fault type to obtain the corresponding exception handling strategy, for example, if the target fault type is component A is damaged, the exception handling strategy is to replace component A. After obtaining the exception handling strategy, the equipment maintenance module 203 sends the exception handling strategy and the maintenance request to the maintenance personnel, so that the maintenance personnel can clearly know the location information corresponding to the target user and the maintenance tools that need to be carried when repairing the target device, thereby avoiding the failure to carry out maintenance in time due to the failure to carry the corresponding parts, resulting in low user satisfaction.

[0038] For example, by centrally managing the smart devices of multiple apartments (such as smart door locks, water and electricity meters, washing machines, air conditioners, etc.), the centralized apartment management system 200 can achieve unified monitoring and maintenance of equipment. All equipment status and maintenance information are managed in a centralized platform, and fault information and maintenance requests can also be assigned to corresponding maintenance personnel through the unified platform. This centralized management makes maintenance more efficient and resource allocation more reasonable. When a device failure occurs, the centralized apartment management system 200 generates an exception handling strategy based on the centralized target operation data, and sends these exception handling strategies together with the maintenance request to the relevant maintenance personnel. Centralized fault handling ensures that faults of all equipment in the apartment can be detected and handled through the same platform, improving the speed and consistency of system response.

[0039] In some implementations, the data analysis module, during the process of identifying anomalies in the initial operation data to obtain target operation data corresponding to the target device, performs:

[0040] Determine a first segmentation window and a target step size, and perform data segmentation on the initial operation data according to the first segmentation window and the target step size to obtain first target window data;

[0041] Obtaining any two window data from the first target window data and determining them as first window data and second window data;

[0042] Subtracting the first window data from the second window data in order and taking the maximum absolute value to obtain a first window distance between the first window data and the second window data;

[0043] Determine a first entropy value corresponding to the initial running data under the first segmentation window according to the first window distance;

[0044] Determine a second segmentation window, and perform data segmentation on the initial operation data according to the second segmentation window and the target step size to obtain second target window data;

[0045] Obtaining any two window data from the second target window data as third window data and fourth window data;

[0046] Subtracting the third window data from the fourth window data in sequence and taking the maximum absolute value to obtain a second window distance between the third window data and the fourth window data;

[0047] Determine a second entropy value corresponding to the initial running data in the second segmentation window according to the second window distance;

[0048] Anomalies are identified on the initial operation data according to the first entropy value and the second entropy value to obtain the target operation data corresponding to the target device.

[0049] In some implementations, the data analysis module, in the process of determining a first entropy value corresponding to the initial running data in the first segmentation window according to the first window distance and determining a second entropy value corresponding to the initial running data in the second segmentation window according to the second window distance, performs:

[0050] Determine a target threshold, and count a first quantity corresponding to when the first window distance is less than the target threshold;

[0051] Determine a first target quantity corresponding to the first target window data;

[0052] Calculating a ratio between the first quantity and the first target quantity to obtain a first ratio, and obtaining the first entropy value according to the first ratio;

[0053] Counting a second number corresponding to when the second window distance is less than the target threshold;

[0054] Determine the second target quantity corresponding to the second target window data;

[0055] Calculating a ratio between the second quantity and the second target quantity to obtain a second ratio, and obtaining the second entropy value according to the second ratio;

[0056] The first entropy value and the second entropy value are obtained according to the following formula:

[0057] ;

[0058] ;

[0059] in, represents the first entropy value, sum1 represents the first target quantity, ln represents taking the natural logarithm, represents the first quantity corresponding to the i-th first target window data; represents the second entropy value, sum2 represents the second target quantity, Represents the second quantity corresponding to the j-th second target window data.

[0060] In some implementations, the equipment maintenance module, during the process of extracting abnormal features from the target operation data to obtain a corresponding target feature vector, performs:

[0061] Performing data statistical analysis on the target operation data to obtain a statistical feature vector corresponding to the target operation data;

[0062] Performing time-frequency analysis on the target operation data to obtain a time-frequency feature vector corresponding to the target operation data;

[0063] The statistical feature vector and the time-frequency feature vector are vector-fused to obtain the target feature vector.

[0064] In some implementations, the equipment maintenance module, in the process of obtaining the target fault type corresponding to the target equipment by performing abnormal classification according to the target feature vector, performs:

[0065] Obtaining a fault classification model corresponding to the target device from the database according to the target identifier;

[0066] Performing abnormal classification on the target feature vector according to the fault classification model to obtain a fault distribution result corresponding to the target operation data;

[0067] The target fault type corresponding to the target device is determined according to the fault distribution result.

[0068] In some implementations, the equipment maintenance module, in the process of performing abnormal classification on the target feature vector according to the fault classification model to obtain the fault distribution result corresponding to the target operation data, performs:

[0069] Obtaining hidden variable information corresponding to the target feature vector using a long short-term memory network of the fault classification model;

[0070] Determine the fusion weight corresponding to the target feature vector according to the hidden variable information using the self-attention network of the fault classification model;

[0071] Using the vector fusion network of the fault classification model to perform feature fusion on the target feature vector according to the fusion weight to obtain a fused feature vector;

[0072] The fused feature vector is classified and identified according to the fault classification network of the fault classification model to obtain the fault distribution result corresponding to the target operation data.

[0073] In some implementations, after sending the exception handling strategy and the maintenance request to the maintenance personnel, the equipment maintenance module further performs:

[0074] Obtain a maintenance processing request corresponding to the maintenance personnel, and send the maintenance processing request to a target terminal associated with the target user, so that after the target terminal receives the maintenance processing request, the target user determines the maintenance time according to the target terminal so that the maintenance personnel processes the maintenance request according to the maintenance time and the exception handling strategy.

[0075] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the centralized apartment management system 200 described above can refer to the corresponding process in the following centralized apartment management method embodiment, and will not be repeated here.

[0076] Please refer to Figure 2 , Figure 2 A flowchart of a centralized apartment management method provided by an embodiment of the present invention.

[0077] like Figure 2 As shown, the centralized apartment management method includes steps S101 to S103.

[0078] Step S101: Obtain a maintenance request corresponding to a target user, and parse the maintenance request to obtain a target device and a target identifier corresponding to the target device.

[0079] Exemplarily, the target user has a client installed in the mobile terminal, and thus sends a maintenance request through the client to obtain the maintenance request corresponding to the target user, and then parses the maintenance request to obtain the target device that needs to be inspected or repaired, and then obtains the user identifier corresponding to the target user from the maintenance request, and then queries the database according to the user identifier and the device name corresponding to the target device to obtain the target identifier corresponding to the target device.

[0080] Step S102: obtaining initial operation data corresponding to the target device from a database according to the target identifier, and performing abnormality identification on the initial operation data to obtain target operation data corresponding to the target device.

[0081] For example, the operation data corresponding to each device in the apartment is collected through sensors and processed in a centralized data platform. By centralizing the operation data of all devices in one database, comprehensive analysis and anomaly detection can be performed. Centralized data processing can not only improve data processing efficiency, but also query the initial operation data corresponding to the target device in a timely manner.

[0082] Exemplarily, data query is performed from the database according to the target identifier to obtain the initial operation data corresponding to the target device. The initial operation data is then cleaned and preprocessed to remove noise and invalid data to ensure the integrity and consistency of the data. Then, the preprocessed data is analyzed using statistical analysis methods or machine learning models to identify abnormal data in the initial operation data, that is, to obtain the target operation data.

[0083] For example, you can choose commonly used anomaly detection algorithms, such as statistical methods (such as Z-score, 3σ principle), time series analysis (such as ARIMA model), cluster analysis (such as K-means), or deep learning methods (such as autoencoder, LSTM), etc., and select a suitable model according to the characteristics of the target device. The preprocessed data is then input into the established model, and the abnormal data points are identified, thereby obtaining the target operation data corresponding to the target device.

[0084] For example, select features that are closely related to the performance of the target device from the initial operating data. These features should be able to reflect the difference between normal and abnormal states. For example, you can focus on statistical properties such as the mean, variance, extreme values, and rate of change of the data. For in-depth analysis of the selected features, some features may need to be transformed (such as normalized, standardized) to adapt to subsequent classification algorithms. Methods such as principal component analysis (PCA) can be used to reduce the data dimension while retaining most of the information. The extracted abnormal features are integrated into a target feature vector.

[0085] In some embodiments, the abnormality identification of the initial operation data to obtain the target operation data corresponding to the target device includes: determining a first segmentation window and a target step size, and performing data segmentation on the initial operation data according to the first segmentation window and the target step size to obtain first target window data; arbitrarily obtaining two window data from the first target window data to determine as first window data and second window data; subtracting the first window data and the second window data in sequence and taking the maximum absolute value to obtain a first window distance between the first window data and the second window data; determining a first entropy value corresponding to the initial operation data under the first segmentation window according to the first window distance. ; Determine a second segmentation window, and perform data segmentation on the initial operation data according to the second segmentation window and the target step size to obtain second target window data; arbitrarily obtain two window data from the second target window data and determine them as third window data and fourth window data; subtract the third window data and the fourth window data in sequence and take the maximum absolute value to obtain a second window distance between the third window data and the fourth window data; determine a second entropy value corresponding to the initial operation data under the second segmentation window according to the second window distance; perform abnormal identification on the initial operation data according to the first entropy value and the second entropy value to obtain the target operation data corresponding to the target device.

[0086] Exemplarily, the first segmentation window and the target step size are determined based on expert experience or historical experience. For example, the size of the first segmentation window is 5 and the target step size is 1. Then, the initial running data is segmented according to the first segmentation window and the target step size to obtain the first target window data. Assuming that the initial running data contains 10 data, the first target window data contains 6 window data.

[0087] Exemplarily, any two window data are obtained from the first target window data and determined as the first window data and the second window data, and then the first window data and the second window data are subtracted in sequence to obtain multiple first subtraction results, and then after taking the absolute values ​​of all the first subtraction results, the maximum value after taking the absolute values ​​is obtained, and then the maximum value is determined as the first window distance between the first window data and the second window data.

[0088] For example, the first window data includes x1, x2, x3, x4, and x5, and the second window data includes x2, x3, x4, x5, and x6. Then, x1 in the first window data and x2 in the second window data, x2 in the first window data and x3 in the second window data, x3 in the first window data and x4 in the second window data, x4 in the first window data and x5 in the second window data, and x5 in the first window data and x6 in the second window data are sequentially subtracted to obtain first subtraction results diff1, diff2, diff3, diff4, and diff5, so that Take the absolute values ​​of the first subtraction results diff1, diff2, diff3, diff4, and diff5 to obtain abs(diff1), abs(diff2), abs(diff3), abs(diff4), and abs(diff5), where abs means taking the absolute value, and then determine the result of max(abs(diff1), abs(diff2), abs(diff3), abs(diff4), abs(diff5)) as the first window distance between the first window data and the second window data, where max means taking the maximum value.

[0089] Exemplarily, the above steps are performed on any two window data in the first target window data, so as to obtain the first window distance corresponding to the any two window data.

[0090] Exemplarily, the frequency of occurrence of each first window distance is counted, and then the first probability corresponding to the first window distance is determined based on the frequency, so as to calculate the entropy value based on the first probability to obtain the first entropy value corresponding to the initial running data in the first segmentation window.

[0091] For example, the first target window data contains 6 window data, so the distance calculation of any two window data can generate 15 first window distances. This combination can be calculated by the "combination formula" in combinatorial mathematics, that is, C(6,2) = 15. Then count the frequency of occurrence of these 15 first window distances. By counting the number of times each specific distance appears, the frequency distribution of each first window distance can be obtained. This process can not only provide the absolute number of occurrences of each distance, but also reveal the relative distribution between distances. Using these frequency data, the corresponding probability of each first window distance can be further calculated. Specifically, the probability of a first window distance can be expressed as the ratio of the frequency of occurrence of the distance to the total frequency of all distances. This probability reflects the relative importance or prevalence of the first window distance in the overall data. After obtaining the probabilities of all first window distances, entropy calculations can be performed based on these probabilities. Entropy is an important indicator for measuring uncertainty or information volume. The higher the entropy value, the more complex the system state and the richer the information volume. By substituting the probability of each first window distance into the entropy formula, the first entropy value corresponding to the initial running data under the first segmentation window can be calculated.

[0092] Exemplarily, a split window with a size different from that of the first split window is determined as the second split window, and the initial running data is split according to the second split window and the target step size to obtain the second target window data, wherein, assuming that the initial running data contains 10 data and the second split window is 6, the second target window data contains 5 window data.

[0093] Exemplarily, any two window data are obtained from the second target window data and determined as the third window data and the fourth window data, and then the third window data and the fourth window data are subtracted in sequence to obtain multiple second subtraction results, and then after taking the absolute values ​​of all the second subtraction results, the maximum value after taking the absolute values ​​is obtained, and then the maximum value is determined as the fourth window distance between the third window data and the fourth window data.

[0094] For example, the third window data includes x1, x2, x3, x4, x5, and x6, and the fourth window data includes x2, x3, x4, x5, x6, and x7. Then, x1 in the third window data and x2 in the fourth window data, x2 in the third window data and x3 in the fourth window data, x3 in the third window data and x4 in the fourth window data, x4 in the third window data and x5 in the fourth window data, x5 in the third window data and x6 in the fourth window data, and x6 in the third window data and x7 in the fourth window data are subtracted in sequence to obtain the second subtraction results diff1, diff2, diff3, diff4, diff5, diff 6, thereby taking the absolute values ​​of the second subtraction results diff1, diff2, diff3, diff4, diff5, and diff6 to obtain abs(diff1), abs(diff2), abs(diff3), abs(diff4), abs(diff5), and abs(diff6), where abs means taking the absolute value, and then determining the result of max(abs(diff1), abs(diff2), abs(diff3), abs(diff4), abs(diff5), and abs(diff6)) as the second window distance between the third window data and the fourth window data, where max means taking the maximum value.

[0095] Exemplarily, the above steps are performed on any two window data in the second target window data, so as to obtain the second window distance corresponding to the any two window data.

[0096] Exemplarily, the frequency of occurrence of each second window distance is counted, and then the second probability corresponding to the second window distance is determined based on the frequency, so as to calculate the entropy value based on the second probability to obtain the second entropy value corresponding to the initial running data in the second segmentation window.

[0097] For example, the second target window data contains 5 window data, so the distance calculation of any two window data can generate 10 first window distances. This combination can be calculated by the "combination formula" in combinatorial mathematics, that is, C(5,2) = 10. Then count the frequencies of these 10 second window distances. By counting the number of times each specific distance appears, the frequency distribution of each second window distance can be obtained. This process can not only provide the absolute number of occurrences of each distance, but also reveal the relative distribution between distances. Using these frequency data, the corresponding probability of each second window distance can be further calculated. Specifically, the probability of a second window distance can be expressed as the ratio of the frequency of occurrence of the distance to the total frequency of all distances. This probability reflects the relative importance or prevalence of the second window distance in the overall data. After obtaining the probabilities of all second window distances, entropy calculations can be performed based on these probabilities. Entropy is an important indicator for measuring uncertainty or information volume. The higher the entropy value, the more complex the system state and the richer the information volume. By substituting the probability of each second window distance into the entropy formula, the second entropy value corresponding to the initial running data under the second segmentation window can be calculated.

[0098] Exemplarily, the difference between the first entropy value and the second entropy value is calculated. When the difference is greater than or equal to a preset value, it is determined that the data fluctuations in the initial operating data are large, and the initial operating data are determined to be the target operating data; when the difference is less than the preset value, it is determined that the data fluctuations in the initial operating data are small, and the data of the next time period corresponding to the initial operating data is determined as the initial operating data, and the above steps are repeated to identify anomalies in the initial operating data to obtain all the target operating data.

[0099] Specifically, different segmentation windows can capture the changes in data at different window scales, thereby improving the accuracy and robustness of anomaly detection. Anomaly detection can be performed through the changes in the entropy values ​​of the initial running data at different window scales, which can effectively improve the efficiency and accuracy of data processing and provide good support for the accurate identification of subsequent fault types.

[0100] In some embodiments, determining a first entropy value corresponding to the initial running data under the first segmentation window according to the first window distance and determining a second entropy value corresponding to the initial running data under the second segmentation window according to the second window distance includes: determining a target threshold, and counting a first quantity corresponding to when the first window distance is less than the target threshold; determining a first target quantity corresponding to the first target window data; calculating a ratio between the first quantity and the first target quantity to obtain a first ratio, and obtaining the first entropy value according to the first ratio; counting a second quantity corresponding to when the second window distance is less than the target threshold; determining a second target quantity corresponding to the second target window data; calculating a ratio between the second quantity and the second target quantity to obtain a second ratio, and obtaining the second entropy value according to the second ratio; wherein the first entropy value and the second entropy value are obtained according to the following formulas:

[0101] ;

[0102] ;

[0103] in, represents the first entropy value, sum1 represents the first target quantity, ln represents taking the natural logarithm, represents the first quantity corresponding to the i-th first target window data; represents the second entropy value, sum2 represents the second target quantity, Represents the second quantity corresponding to the j-th second target window data.

[0104] Exemplarily, the target threshold is determined, and the first window distance is the distance information between any two window data in the first target window data. Then, a window data is arbitrarily determined from the first target window data, and then the first window distance between the determined window data and all the remaining window data is obtained, and then the first window distance is compared with the target threshold respectively, so as to obtain the first quantity corresponding to when the first window distance is less than the target threshold. The first quantity is used to characterize the quantity of the first window distance between any window data in the first target window data and the remaining window data is less than the target threshold, so as to calculate the ratio of the first quantity and the first target quantity to obtain the first ratio, and then obtain the first quantity corresponding to the first window distance between each window data in the first target window data and the remaining window data, so as to obtain the first ratio corresponding to each window data in the first target window data according to the first quantity. Finally, the first entropy value is obtained according to the following formula:

[0105] ;

[0106] in, represents the first entropy value, sum1 represents the first target quantity, ln represents taking the natural logarithm, Represents the first quantity corresponding to the i-th first target window data.

[0107] Exemplarily, the second window distance is the distance information between any two window data in the second target window data, then a window data is arbitrarily determined from the second target window data, and then the second window distance between the determined window data and all the remaining window data is obtained, and then the second window distance is compared with the target threshold respectively, so as to obtain the second number corresponding to when the second window distance is less than the target threshold, and the second number is used to characterize the number of times that the second window distance between any window data in the second target window data and the remaining window data is less than the target threshold, so as to calculate the ratio of the second number and the second target number to obtain the second ratio, and then obtain the second number corresponding to the second window distance between each window data in the second target window data and the remaining window data, so as to obtain the second ratio corresponding to each window data in the second target window data according to the second number. Finally, the second entropy value is obtained according to the following formula:

[0108] ;

[0109] in, represents the second entropy value, sum2 represents the second target number, ln represents the natural logarithm, Indicates the second quantity corresponding to the j-th second target window data.

[0110] Specifically, the ratio reflects the proportion of data that meets the threshold conditions in the overall data. By calculating the ratio, we can measure the uniformity of data distribution, thereby evaluating the complexity and uncertainty of the data. Furthermore, by calculating the ratio and obtaining the entropy value based on the ratio, we can more accurately reflect the complexity and changing trend of the data under different windows. This is very useful for further data analysis and anomaly detection, and can effectively improve the accuracy of data screening, enhance the understanding of data complexity and changing trends, and provide good support for subsequent abnormal data detection.

[0111] Step S103, extract abnormal features from the target operating data to obtain a corresponding target feature vector, and perform abnormal classification based on the target feature vector to obtain a target fault type corresponding to the target device; determine an abnormality handling strategy corresponding to the maintenance request based on the target fault type; and send the abnormality handling strategy and the maintenance request to maintenance personnel, so that the maintenance personnel perform maintenance processing on the maintenance request according to the abnormality handling strategy.

[0112] For example, select features that are closely related to the performance of the target device from the initial operating data. These features should be able to reflect the difference between normal and abnormal states. For example, you can focus on statistical properties such as the mean, variance, extreme values, and rate of change of the data. For in-depth analysis of the selected features, some features may need to be transformed (such as normalized, standardized) to adapt to subsequent classification algorithms. Methods such as principal component analysis (PCA) can be used to reduce the data dimension while retaining most of the information. The extracted abnormal features are integrated into a target feature vector.

[0113] For example, according to the characteristics of the feature vector, a suitable machine learning or statistical classification algorithm is selected, such as support vector machine (SVM), decision tree, random forest, neural network, etc. The classification model is trained using a labeled feature vector set to establish a mapping relationship between features and fault types. Thus, the target feature vector is input into the trained model for abnormal classification, and the corresponding target fault type is output.

[0114] Exemplarily, the target fault type is queried in the policy mapping table according to the policy mapping table to obtain the corresponding exception handling strategy, for example: if the target fault type is component A is damaged, the exception handling strategy is to replace component A. After obtaining the exception handling strategy, the exception handling strategy and the maintenance request are sent to the maintenance personnel, so that the maintenance personnel can clearly know the location information corresponding to the target user and the maintenance tools that need to be carried when repairing the target equipment, thereby avoiding the failure to carry out maintenance in time due to the failure to carry the corresponding parts, resulting in low user satisfaction.

[0115] In some embodiments, the abnormal feature extraction of the target operation data to obtain the corresponding target feature vector includes: performing data statistical analysis on the target operation data to obtain the statistical feature vector corresponding to the target operation data; performing time-frequency analysis on the target operation data to obtain the time-frequency feature vector corresponding to the target operation data; and performing vector fusion of the statistical feature vector and the time-frequency feature vector to obtain the target feature vector.

[0116] Exemplarily, the statistical indicators of the analysis are determined, such as mean, variance, maximum value, minimum value, median, etc. These indicators can reflect the central tendency and dispersion degree of the data. The target operation data is batch processed, the selected statistical indicators are calculated one by one, and the calculation results of these indicators are combined into a statistical feature vector.

[0117] Exemplarily, a time-frequency analysis method is determined. Common time-frequency analysis methods include Fourier transform, wavelet transform, etc., and an analysis method suitable for the characteristics of the target operation data is selected. The target operation data is input into the selected time-frequency analysis method to generate a time-frequency spectrum. Key features such as the peak value, energy distribution, frequency range, etc. of the spectrum are extracted from the generated time-frequency spectrum to form a time-frequency feature vector.

[0118] Exemplarily, ensure that the dimensions of the statistical feature vector and the time-frequency feature vector are aligned for effective fusion. According to the determined fusion algorithm, the statistical feature vector and the time-frequency feature vector are then fused to generate a target feature vector. The fusion algorithm includes but is not limited to weighted fusion, series fusion, etc.

[0119] Specifically, a variety of features are extracted from the target operation data, and the statistical characteristics and time-frequency characteristics of the data are comprehensively considered through vector fusion, so as to obtain a more representative and discriminative target feature vector, providing strong support for subsequent analysis and decision-making.

[0120] In some embodiments, performing abnormal classification according to the target feature vector to obtain a target fault type corresponding to the target device includes: obtaining a fault classification model corresponding to the target device from the database according to the target identifier; performing abnormal classification on the target feature vector according to the fault classification model to obtain a fault distribution result corresponding to the target operating data; and determining the target fault type corresponding to the target device according to the fault distribution result.

[0121] Exemplarily, a query operation is first performed according to a target identifier such as a device ID to obtain a fault classification model corresponding to the target device from a database. The fault classification model may be a machine learning model trained based on historical data or a set of rules. After obtaining the fault classification model, it is necessary to confirm the integrity and authenticity of the model, including checking the version, training date, accuracy, etc. of the fault classification model to ensure that it is suitable for the current fault diagnosis task.

[0122] Exemplarily, ensure that the format of the target feature vector matches the input format required by the fault classification model, and then input the target feature vector into the acquired fault classification model to perform the classification operation. The fault classification model will output the fault distribution results of the target operating data. The fault distribution results include the probability distribution of different fault types, the labels of each fault class, and the corresponding confidence scores.

[0123] Exemplarily, the fault distribution results are analyzed, and corresponding decision criteria are set according to business requirements and the degree of fault impact to determine which fault type in the fault distribution results is considered as the target fault type. For example, a threshold is set, and when the confidence or probability of a certain fault type exceeds the threshold, it is identified as the corresponding target fault type. The target fault type can be one or more.

[0124] In some embodiments, the target feature vector is classified as abnormal according to the fault classification model to obtain the fault distribution result corresponding to the target operation data, including: using the long short-term memory network of the fault classification model to obtain the hidden variable information corresponding to the target feature vector; using the self-attention network of the fault classification model to determine the fusion weight corresponding to the target feature vector according to the hidden variable information; using the vector fusion network of the fault classification model to perform feature fusion on the target feature vector according to the fusion weight to obtain a fused feature vector; and classifying and identifying the fused feature vector according to the fault classification network of the fault classification model to obtain the fault distribution result corresponding to the target operation data.

[0125] Exemplarily, the fault classification model includes a long short-term memory network, a self-attention network, a vector fusion network, and a fault classification network.

[0126] Exemplarily, ensure that the target feature vector has been properly preprocessed, including normalization, denoising, etc., to meet the input requirements of the long short-term memory network. In the fault classification model, the long short-term memory network is used to capture the long-term dependencies in the sequence data. The long short-term memory network is trained using historical fault data and its labels so that it can learn the mapping from feature sequence to hidden state. The target feature vector is then forward propagated through the trained long short-term memory network to obtain the hidden state of each time step. These hidden states constitute the hidden variable information, which encodes the temporal dependencies in the feature vector.

[0127] For example, the long short-term memory network can effectively capture the long-term dependencies in time series data. By obtaining hidden variable information, the important features of the target feature vector in the time dimension can be extracted to enhance the understanding of the time series of the data. This helps to better identify the dynamic changes of the equipment before and after the failure, thereby improving the accuracy of fault prediction.

[0128] For example, a self-attention mechanism is integrated into the fault classification model, which can calculate the correlation between hidden states at different time steps. The self-attention network is usually composed of a multi-head attention mechanism to capture different dependencies in the input sequence. The hidden variable information output by the long short-term memory network is used as the input of the self-attention network, and the fusion weights of each hidden state are obtained by calculating the attention score and applying the softmax function. These weights reflect the importance of features at different time steps.

[0129] For example, according to the fusion weights calculated by the self-attention network, the hidden states output by the long short-term memory network are weighted summed to obtain a fused feature vector that incorporates temporal information. This step is essentially a weighted aggregation of features at different time steps. After feature fusion, it is also necessary to add nonlinear activation functions, perform layer normalization, etc. to enhance the expressive power of the fused feature vector.

[0130] Exemplarily, by calculating the fusion weights based on the hidden variable information, it is possible to adaptively focus on important features. This means that the system can automatically identify the features that have the greatest impact on fault classification and improve the model's ability to respond to different fault conditions. By dynamically adjusting the weights, the model can more flexibly process data in various scenarios and reduce the impact of noise features. The target feature vectors are weighted and fused using the calculated fusion weights, which helps integrate information from different sources to produce a richer and more representative fused feature vector. The fused feature vector not only retains important timing information, but also reflects the relationship between multiple features, improving the expressiveness of the features. Training the fault classification network using the fused feature vector can significantly improve the accuracy and robustness of fault classification.

[0131] Exemplarily, a fault classification network is constructed in the last part of the fault classification model, which can receive the fused feature vector as input and output the fault distribution result. For example, the fault classification network is a structure of a fully connected layer plus a softmax activation function, which is used to map the fused features to the probability distribution of the fault category. Thus, the fused feature vector is input into the fault classification network to obtain the fault distribution result corresponding to the target operation data. This result is a probability distribution that reflects the possibility of different fault types.

[0132] Specifically, by combining time series features and dynamic weighting of the self-attention mechanism, the model is able to classify more effectively in complex fault modes, which in turn enables the model to better cope with the uncertainty and diversity in equipment operation and improve the reliability of fault detection.

[0133] In some embodiments, after sending the exception handling strategy and the maintenance request to the maintenance personnel, the method further includes: obtaining the maintenance processing request corresponding to the maintenance personnel, and sending the maintenance processing request to the target terminal associated with the target user, so that after the target terminal receives the maintenance processing request, the target user determines the maintenance time according to the target terminal so that the maintenance personnel processes the maintenance request according to the maintenance time and the exception handling strategy.

[0134] For example, the maintenance personnel determine the maintenance plan corresponding to the target device in the maintenance request based on the exception handling strategy and the maintenance request, and then search the database for the target user information related to the target device and its corresponding target terminal according to the maintenance request, which may include the user's contact information, the registered terminal device or the unique identifier of the application, and then send the maintenance processing request to the target terminal associated with the target user. The maintenance processing request is encapsulated into a message format, which contains all necessary information, such as request content, requester identity, device information, maintenance time, etc.

[0135] Exemplarily, after the target terminal receives the maintenance processing request, the target user will receive a notification, which may be a pop-up window of the application, a sound reminder, or other forms of prompts. Then, after the target user receives the maintenance processing request, the target user selects a suitable maintenance time for maintenance according to his or her schedule and the urgency of the equipment. Once the target user sets the maintenance time, this information is fed back to the maintenance personnel in a timely manner. The maintenance personnel can be notified by notifying or updating the maintenance request status. Finally, based on the set maintenance time and exception handling strategy, the maintenance personnel carry out subsequent maintenance preparations, tool preparations, and on-site arrangements in accordance with the process. In this way, the maintenance personnel's requests can be processed efficiently, and it is ensured that the target user can arrange the maintenance time in a timely manner, thereby improving the efficiency of the maintenance process and user satisfaction.

[0136] The embodiment of the present invention provides a centralized apartment management method, which includes obtaining a maintenance request corresponding to a target user, parsing the maintenance request to obtain a target device and a target identifier corresponding to the target device, and then obtaining the initial operation data corresponding to the target device from a database according to the target identifier, and then performing abnormal identification on the initial operation data to obtain the target operation data corresponding to the target device, thereby performing abnormal feature extraction on the target operation data to obtain the corresponding target feature vector, and performing abnormal classification according to the target feature vector to obtain the target fault type corresponding to the target device; determining the abnormal handling strategy corresponding to the maintenance request according to the target fault type; sending the abnormal handling strategy and the maintenance request to the maintenance personnel, so that the maintenance personnel perform maintenance processing on the maintenance request according to the abnormal handling strategy. The method performs abnormal identification on the initial operation data, generates the target operation data, further improves the accuracy of fault judgment, and then performs abnormal feature extraction on the target operation data to form a target feature vector, which is helpful to quickly identify and classify the abnormal state of the equipment. Perform abnormal classification according to the target feature vector, determine the target fault type of the target device, and realize intelligent diagnosis of the fault. According to the target fault type, the optimal abnormal handling strategy can be generated to reduce the deviation and delay of human judgment. Thus, the exception handling strategy and maintenance request are sent directly to the maintenance personnel, ensuring that they can quickly and accurately perform maintenance and improve service efficiency. This significantly improves the efficiency and accuracy of apartment equipment maintenance, optimizes the user demand response process, reduces fault handling time, and improves customer satisfaction. This solves the problem of low efficiency and insufficient accuracy in the process of apartment equipment maintenance and user demand response in related technologies.

[0137] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention.

[0138] like Figure 3 As shown, the terminal device 300 includes a processor 301 and a memory 302 , and the processor 301 and the memory 302 are connected via a bus 303 , such as an I2C (Inter-integrated Circuit) bus.

[0139] Specifically, the processor 301 is used to provide computing and control capabilities to support the operation of the entire terminal device. The processor 301 can be a central processing unit (CPU), and the processor 301 can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0140] Specifically, the memory 302 may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.

[0141] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the embodiment of the present invention, and does not constitute a limitation on the terminal device to which the embodiment of the present invention is applied. The specific server may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0142] The processor is used to run a computer program stored in the memory, and implement any one of the centralized apartment management methods provided by the embodiments of the present invention when executing the computer program.

[0143] In one embodiment, the processor is used to run a computer program stored in the memory, and implements the following steps when executing the computer program:

[0144] Obtaining a maintenance request corresponding to a target user, and parsing the maintenance request to obtain a target device and a target identifier corresponding to the target device;

[0145] Obtaining initial operation data corresponding to the target device from a database according to the target identifier, and performing abnormality identification on the initial operation data to obtain target operation data corresponding to the target device;

[0146] Perform abnormal feature extraction on the target operation data to obtain a corresponding target feature vector, and perform abnormal classification based on the target feature vector to obtain a target fault type corresponding to the target equipment; determine an abnormal handling strategy corresponding to the maintenance request based on the target fault type; send the abnormal handling strategy and the maintenance request to maintenance personnel, so that the maintenance personnel perform maintenance processing on the maintenance request according to the abnormal handling strategy.

[0147] In some implementations, the processor 301, during the process of identifying anomalies in the initial operation data to obtain target operation data corresponding to the target device, executes:

[0148] Determine a first segmentation window and a target step size, and perform data segmentation on the initial operation data according to the first segmentation window and the target step size to obtain first target window data;

[0149] Obtaining any two window data from the first target window data and determining them as first window data and second window data;

[0150] Subtracting the first window data from the second window data in order and taking the maximum absolute value to obtain a first window distance between the first window data and the second window data;

[0151] Determine a first entropy value corresponding to the initial running data under the first segmentation window according to the first window distance;

[0152] Determine a second segmentation window, and perform data segmentation on the initial operation data according to the second segmentation window and the target step size to obtain second target window data;

[0153] Obtaining any two window data from the second target window data as third window data and fourth window data;

[0154] Subtracting the third window data from the fourth window data in sequence and taking the maximum absolute value to obtain a second window distance between the third window data and the fourth window data;

[0155] Determine a second entropy value corresponding to the initial running data in the second segmentation window according to the second window distance;

[0156] Anomalies are identified on the initial operation data according to the first entropy value and the second entropy value to obtain the target operation data corresponding to the target device.

[0157] In some implementations, the processor 301, in the process of determining a first entropy value corresponding to the initial running data in the first segmentation window according to the first window distance and determining a second entropy value corresponding to the initial running data in the second segmentation window according to the second window distance, executes:

[0158] Determine a target threshold, and count a first quantity corresponding to when the first window distance is less than the target threshold;

[0159] Determine a first target quantity corresponding to the first target window data;

[0160] Calculating a ratio between the first quantity and the first target quantity to obtain a first ratio, and obtaining the first entropy value according to the first ratio;

[0161] Counting a second number corresponding to when the second window distance is less than the target threshold;

[0162] Determine the second target quantity corresponding to the second target window data;

[0163] Calculating a ratio between the second quantity and the second target quantity to obtain a second ratio, and obtaining the second entropy value according to the second ratio;

[0164] The first entropy value and the second entropy value are obtained according to the following formula:

[0165] ;

[0166] ;

[0167] in, represents the first entropy value, sum1 represents the first target quantity, ln represents taking the natural logarithm, represents the first quantity corresponding to the i-th first target window data; represents the second entropy value, sum2 represents the second target quantity, Represents the second quantity corresponding to the j-th second target window data.

[0168] In some implementations, the processor 301, during the process of extracting abnormal features from the target operating data to obtain a corresponding target feature vector, executes:

[0169] Performing data statistical analysis on the target operation data to obtain a statistical feature vector corresponding to the target operation data;

[0170] Performing time-frequency analysis on the target operation data to obtain a time-frequency feature vector corresponding to the target operation data;

[0171] The statistical feature vector and the time-frequency feature vector are vector-fused to obtain the target feature vector.

[0172] In some implementations, during the process of performing abnormal classification according to the target feature vector to obtain a target fault type corresponding to the target device, the processor 301 executes:

[0173] Obtaining a fault classification model corresponding to the target device from the database according to the target identifier;

[0174] Performing abnormal classification on the target feature vector according to the fault classification model to obtain a fault distribution result corresponding to the target operation data;

[0175] The target fault type corresponding to the target device is determined according to the fault distribution result.

[0176] In some implementations, the processor 301, in the process of performing abnormal classification on the target feature vector according to the fault classification model to obtain the fault distribution result corresponding to the target operation data, executes:

[0177] Obtaining hidden variable information corresponding to the target feature vector using a long short-term memory network of the fault classification model;

[0178] Determine the fusion weight corresponding to the target feature vector according to the hidden variable information using the self-attention network of the fault classification model;

[0179] Using the vector fusion network of the fault classification model to perform feature fusion on the target feature vector according to the fusion weight to obtain a fused feature vector;

[0180] The fused feature vector is classified and identified according to the fault classification network of the fault classification model to obtain the fault distribution result corresponding to the target operation data.

[0181] In some implementations, after the processor 301 sends the exception handling strategy and the maintenance request to the maintenance personnel, it further executes:

[0182] Obtain a maintenance processing request corresponding to the maintenance personnel, and send the maintenance processing request to a target terminal associated with the target user, so that after the target terminal receives the maintenance processing request, the target user determines the maintenance time according to the target terminal so that the maintenance personnel processes the maintenance request according to the maintenance time and the exception handling strategy.

[0183] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned centralized apartment management method embodiment, and will not be repeated here.

[0184] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any centralized apartment management method provided in the description of the embodiment of the present invention.

[0185] The storage medium may be an internal storage unit of the terminal device described in the foregoing embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc., equipped on the terminal device.

[0186] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0187] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0188] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A centralized apartment management system, characterized in that: The centralized apartment management system includes a customer service module, a data analysis module, and a repair and maintenance module, wherein: The customer service module is used to obtain a maintenance request corresponding to a target user, and parse the maintenance request to obtain a target device and a target identifier corresponding to the target device; The data analysis module is used to obtain the initial operation data corresponding to the target device from the database according to the target identifier, and perform abnormality identification on the initial operation data to obtain the target operation data corresponding to the target device; The equipment maintenance module is used to extract abnormal features from the target operating data to obtain a corresponding target feature vector, and to perform abnormal classification based on the target feature vector to obtain a target fault type corresponding to the target equipment; determine an abnormality handling strategy corresponding to the maintenance request based on the target fault type; and send the abnormality handling strategy and the maintenance request to maintenance personnel, so that the maintenance personnel can perform maintenance processing on the maintenance request based on the abnormality handling strategy.

2. A centralized apartment management method, characterized in that: The method comprises: Obtaining a maintenance request corresponding to a target user, and parsing the maintenance request to obtain a target device and a target identifier corresponding to the target device; Obtaining initial operation data corresponding to the target device from a database according to the target identifier, and performing abnormality identification on the initial operation data to obtain target operation data corresponding to the target device; Perform abnormal feature extraction on the target operation data to obtain a corresponding target feature vector, and perform abnormal classification based on the target feature vector to obtain a target fault type corresponding to the target equipment; determine an abnormal handling strategy corresponding to the maintenance request based on the target fault type; send the abnormal handling strategy and the maintenance request to maintenance personnel, so that the maintenance personnel perform maintenance processing on the maintenance request according to the abnormal handling strategy.

3. The method according to claim 2, characterized in that The performing abnormal identification on the initial operation data to obtain target operation data corresponding to the target device includes: Determine a first segmentation window and a target step size, and perform data segmentation on the initial operation data according to the first segmentation window and the target step size to obtain first target window data; Obtaining any two window data from the first target window data and determining them as first window data and second window data; Subtracting the first window data from the second window data in order and taking the maximum absolute value to obtain a first window distance between the first window data and the second window data; Determine a first entropy value corresponding to the initial running data under the first segmentation window according to the first window distance; Determine a second segmentation window, and perform data segmentation on the initial operation data according to the second segmentation window and the target step size to obtain second target window data; Obtaining any two window data from the second target window data as third window data and fourth window data; Subtracting the third window data from the fourth window data in sequence and taking the maximum absolute value to obtain a second window distance between the third window data and the fourth window data; Determine a second entropy value corresponding to the initial running data in the second segmentation window according to the second window distance; Anomalies are identified on the initial operation data according to the first entropy value and the second entropy value to obtain the target operation data corresponding to the target device.

4. The method according to claim 3, characterized in that Determining a first entropy value corresponding to the initial running data in the first segmentation window according to the first window distance and determining a second entropy value corresponding to the initial running data in the second segmentation window according to the second window distance includes: Determine a target threshold, and count a first quantity corresponding to when the first window distance is less than the target threshold; Determine a first target quantity corresponding to the first target window data; Calculating a ratio between the first quantity and the first target quantity to obtain a first ratio, and obtaining the first entropy value according to the first ratio; Counting a second number corresponding to when the second window distance is less than the target threshold; Determine the second target quantity corresponding to the second target window data; Calculating a ratio between the second quantity and the second target quantity to obtain a second ratio, and obtaining the second entropy value according to the second ratio; The first entropy value and the second entropy value are obtained according to the following formula: ； ； in, represents the first entropy value, sum1 represents the first target quantity, ln represents taking the natural logarithm, represents the first quantity corresponding to the i-th data in the first target window; represents the second entropy value, sum2 represents the second target quantity, Represents the second quantity corresponding to the j-th second target window data.

5. The method according to claim 2, characterized in that: The extracting abnormal features of the target operation data to obtain a corresponding target feature vector includes: Performing data statistical analysis on the target operation data to obtain a statistical feature vector corresponding to the target operation data; Performing time-frequency analysis on the target operation data to obtain a time-frequency feature vector corresponding to the target operation data; The statistical feature vector and the time-frequency feature vector are vector-fused to obtain the target feature vector.

6. The method according to claim 2, characterized in that The obtaining a target fault type corresponding to the target device by performing abnormal classification according to the target feature vector includes: Obtaining a fault classification model corresponding to the target device from the database according to the target identifier; Performing abnormal classification on the target feature vector according to the fault classification model to obtain a fault distribution result corresponding to the target operation data; The target fault type corresponding to the target device is determined according to the fault distribution result.

7. The method according to claim 6, characterized in that The performing abnormal classification on the target feature vector according to the fault classification model to obtain a fault distribution result corresponding to the target operation data includes: Obtaining hidden variable information corresponding to the target feature vector using a long short-term memory network of the fault classification model; Determine the fusion weight corresponding to the target feature vector according to the hidden variable information using the self-attention network of the fault classification model; Using the vector fusion network of the fault classification model to perform feature fusion on the target feature vector according to the fusion weight to obtain a fused feature vector; The fused feature vector is classified and identified according to the fault classification network of the fault classification model to obtain the fault distribution result corresponding to the target operation data.

8. The method according to claim 2, characterized in that: After sending the exception handling strategy and the maintenance request to the maintenance personnel, the method further includes: Obtain a maintenance processing request corresponding to the maintenance personnel, and send the maintenance processing request to a target terminal associated with the target user, so that after the target terminal receives the maintenance processing request, the target user determines the maintenance time according to the target terminal so that the maintenance personnel processes the maintenance request according to the maintenance time and the exception handling strategy.

9. A terminal device, characterized in that: The terminal device includes a processor and a memory; The memory is used to store computer programs; The processor is configured to execute the computer program and implement the centralized apartment management method according to any one of claims 2 to 8 when executing the computer program.

10. A computer storage medium for computer storage, characterized in that: The computer storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the centralized apartment management method described in any one of claims 2 to 8.

Images