Equipment data analysis early warning method and system based on Internet of Things, and storage medium

Abstract

The invention discloses an equipment data analysis early warning method and system based on the Internet of Things, and a storage medium, and relates to the technical field of data processing. The method comprises the following steps: acquiring an Internet of Things device, and monitoring a multi-dimensional operation data stream of a target electric meter device in real time; analyzing the data stream to obtain a communication link, and transmitting the data stream to a data processing center based on the communication link; processing the data stream according to a standardized program of a data processing center to obtain a standard multi-dimensional data stream; constructing an electricity meter anomaly analysis adaptive network by using a data mining algorithm, and analyzing a standard data stream to obtain an abnormal operation parameter; a hierarchical early warning mechanism is constructed, early warning is triggered based on abnormal parameters, a strategy is determined, and early warning is performed on the electric meter equipment, so that the technical problem that multi-dimensional data streams and abnormal states of the electric meter equipment cannot be monitored timely and accurately due to insufficient transmission and abnormal detection mechanisms of data of the electric meter equipment is solved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a device data analysis and early warning method, system and storage medium based on the Internet of Things. Background Art

[0002] With the continuous development of smart grid and Internet of Things technologies, electric meters, as important monitoring equipment in power systems, undertake the key task of real-time monitoring of power consumption, operating status and equipment health. However, traditional electric meter equipment monitoring and data transmission methods have many shortcomings. In traditional technologies, the electric meter monitoring system faces multiple challenges in the collection and transmission of electric meter data, including delays in the data transmission process, data loss or incomplete information, which makes it impossible to efficiently process the multi-dimensional data stream of the electric meter equipment, making it difficult for the electric meter monitoring system to accurately monitor the operating status and abnormal performance of the electric meter. In addition, traditional electric meter anomaly detection mostly relies on a single data stream or static analysis method, lacks sufficient intelligent processing capabilities, and cannot effectively deal with the abnormal state of electric meter equipment in complex operating environments. This limitation makes it impossible for traditional monitoring methods to timely and accurately identify faults or potential problems of electric meter equipment, increasing the operating risks of the power system. Summary of the invention

[0003] The present application provides a device data analysis and early warning method, system and storage medium based on the Internet of Things, which solves the technical problem that the multi-dimensional data stream and abnormal state of the electric meter equipment cannot be monitored in a timely and accurate manner due to insufficient transmission and abnormality detection mechanisms of the electric meter equipment data. By integrating the electric meter sensor group and the wireless communication module, and performing abnormal analysis of the data stream based on the electric meter abnormality analysis adaptive network, the technical effect of improving the data transmission efficiency of the electric meter equipment, the abnormality analysis accuracy and the early warning response speed is achieved.

[0004] In view of the above problems, the present application provides a device data analysis and early warning method, system and storage medium based on the Internet of Things.

[0005] The first aspect of the present application provides an IoT-based device data analysis and early warning method, the method comprising: obtaining an IoT device, the IoT device integrating an electric meter sensor group and a wireless communication module, and obtaining a multi-dimensional operation data stream of a target electric meter device through the electric meter sensor group in real time monitoring; performing data transmission analysis on the multi-dimensional operation data stream through the wireless communication module to obtain a multi-level data wireless communication link, and transmitting the multi-dimensional operation data stream to a data processing center based on the multi-level data wireless communication link; obtaining a data standardization program according to the data processing center, and performing standardization processing on the multi-dimensional operation data stream based on the data standardization program to obtain a standard multi-dimensional operation data stream; constructing an electric meter abnormality analysis adaptive network using a data mining algorithm, and performing abnormality analysis on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter; constructing a hierarchical early warning mechanism, triggering the hierarchical early warning mechanism based on the abnormal operation parameters of the electric meter to perform strategy analysis, determine the equipment early warning strategy, and perform operation early warning on the target electric meter device through the equipment early warning strategy.

[0006] The second aspect of the present application provides an IoT-based device data analysis and early warning system, the system comprising: a data monitoring unit: obtaining an IoT device, the IoT device integrating an electric meter sensor group and a wireless communication module, and real-time monitoring and obtaining a multi-dimensional operation data stream of a target electric meter device through the electric meter sensor group; a data transmission unit: performing data transmission and analysis on the multi-dimensional operation data stream through the wireless communication module to obtain a multi-level data wireless communication link, and transmitting the multi-dimensional operation data stream to a data processing center based on the multi-level data wireless communication link; a standardization processing unit: obtaining a data standardization program according to the data processing center, and performing standardization processing on the multi-dimensional operation data stream based on the data standardization program to obtain a standard multi-dimensional operation data stream; an abnormality analysis unit: constructing an electric meter abnormality analysis adaptive network using a data mining algorithm, and performing abnormality analysis on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter; an operation early warning unit: constructing a hierarchical early warning mechanism, triggering the hierarchical early warning mechanism based on the abnormal operation parameters of the electric meter to perform a strategy analysis, determine an equipment early warning strategy, and perform an operation early warning on the target electric meter device through the equipment early warning strategy.

[0007] The third aspect of the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the device data analysis and early warning method based on the Internet of Things provided by the present application.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, an Internet of Things device is obtained, wherein the Internet of Things device integrates an electric meter sensor group and a wireless communication module, and the multi-dimensional operation data stream of the target electric meter device is monitored in real time through the electric meter sensor group; then, the multi-dimensional operation data stream is subjected to data transmission and analysis through the wireless communication module to obtain a multi-level data wireless communication link, and the multi-dimensional operation data stream is transmitted to a data processing center based on the multi-level data wireless communication link; further, a data standardization program is obtained according to the data processing center, and the multi-dimensional operation data stream is subjected to standardization processing based on the data standardization program to obtain a standard multi-dimensional operation data stream; then, an electric meter abnormality analysis adaptive network is constructed using a data mining algorithm, and abnormality analysis is performed on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter; finally, a hierarchical warning mechanism is constructed, and the hierarchical warning mechanism is triggered based on the abnormal operation parameters of the electric meter to perform strategy analysis, determine the equipment warning strategy, and perform operation warning on the target electric meter device through the equipment warning strategy. The technical problem of being unable to timely and accurately monitor the multi-dimensional data stream and abnormal status of the electric meter equipment due to insufficient transmission and anomaly detection mechanism of the electric meter equipment data is solved. By integrating the electric meter sensor group and the wireless communication module, and performing anomaly analysis of the data stream based on the electric meter anomaly analysis adaptive network, the technical effect of improving the data transmission efficiency of the electric meter equipment, the accuracy of anomaly analysis and the speed of early warning response is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a device data analysis and early warning method based on the Internet of Things provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of an IoT-based device data analysis and early warning system provided in an embodiment of the present application.

[0011] Explanation of the accompanying drawings: data monitoring unit 11, data transmission unit 12, standardization processing unit 13, abnormality analysis unit 14, operation warning unit 15. DETAILED DESCRIPTION

[0012] The present application provides a device data analysis and early warning method, system and storage medium based on the Internet of Things, thereby solving the technical problem that the multi-dimensional data stream and abnormal state of the electric meter device cannot be monitored in a timely and accurate manner due to insufficient transmission and abnormality detection mechanisms of the electric meter device data. By integrating the electric meter sensor group and the wireless communication module, and performing abnormal analysis of the data stream based on the electric meter abnormality analysis adaptive network, the technical effect of improving the data transmission efficiency of the electric meter device, the abnormality analysis accuracy and the early warning response speed is achieved.

[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0014] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0015] Embodiment 1, as Figure 1 As shown, the present application provides a device data analysis and early warning method based on the Internet of Things, wherein the method includes: An Internet of Things device is obtained, wherein the Internet of Things device integrates an electric meter sensor group and a wireless communication module, and a multi-dimensional operation data stream of a target electric meter device is obtained by real-time monitoring through the electric meter sensor group.

[0016] In one embodiment, the system terminal obtains an Internet of Things device that integrates an electric meter sensor group and a wireless communication module, and monitors the operation status of the target electric meter device in real time through the integrated electric meter sensor group, collects its multi-dimensional data, such as voltage, current, power, electric meter status data (such as battery status, working status, calibration status) and other related parameters, to form a multi-dimensional operation data stream, providing basic data for subsequent analysis and early warning; the integrated wireless communication module will also transmit the collected multi-dimensional operation data stream to the data processing center by wireless means, perform standardized processing of the multi-dimensional operation data stream, so that subsequent early warnings can be based on more accurate operation data, thereby ensuring the accuracy of the early warning.

[0017] The multi-dimensional operation data stream is subjected to data transmission analysis by the wireless communication module to obtain a multi-level data wireless communication link, and the multi-dimensional operation data stream is transmitted to a data processing center based on the multi-level data wireless communication link.

[0018] In one embodiment, after obtaining the multidimensional operation data stream, the system terminal will parse the multidimensional operation data stream, identify different types of data and communication requirements, and based on these requirements, in combination with the data wireless communication link list of the wireless communication module, distribute the data stream to multiple different communication links to form a multi-level data wireless communication link. These links are hierarchically set according to data transmission requirements, such as communication rate, transmission distance and other factors. Each level of link is responsible for processing a certain amount of data to ensure that data can be transmitted quickly and reliably under different network conditions. Finally, the multidimensional operation data stream will be transmitted to the data processing center through the multi-level data wireless communication link for further processing and analysis.

[0019] Furthermore, the obtaining of a multi-level data wireless communication link includes: Obtain data communication demand factors, wherein the data communication demand factors include communication rate, transmission distance, data type and data sensitivity; analyze the multidimensional operation data flow based on the data communication demand factors to obtain a multidimensional data communication demand parameter set; perform communication diversion on the multidimensional operation data flow according to the multidimensional data communication demand parameter set to obtain a hierarchical communication multidimensional data flow; obtain a data wireless communication link list through the wireless communication module, and perform communication matching with the data wireless communication link list based on the hierarchical communication multidimensional data flow to obtain the multi-level data wireless communication link.

[0020] Preferably, the system terminal obtains key factors affecting data transmission performance according to the specific application scenario of the electric meter device, and these factors include communication rate, transmission distance, data type and data sensitivity, wherein the communication rate determines the amount of data that can be transmitted per second, affecting the transmission speed of the data stream, the transmission distance affects the transmission range of the signal, and long-distance transmission may require a stronger signal or higher bandwidth support, the data type is the category of electric meter data to be transmitted, such as voltage, current, power, temperature, etc., and different categories of data may require different transmission strategies and priority processing, data sensitivity refers to the confidentiality or importance of the data, and some key data may require higher transmission security; then, the system terminal analyzes the multi-dimensional operation data stream of the electric meter device, identifies the specific factor values ​​of the operation data stream of each dimension, and obtains a multi-dimensional data communication requirement parameter set; for the communication rate, the system terminal calculates the required minimum communication rate according to the frequency of change of the operation data stream of each dimension and the size of each data packet, for example, for current and voltage data with higher real-time requirements, a higher communication rate is required, while for periodic The updated meter status data can have a lower rate requirement. By determining the minimum communication rate, it can be ensured that the data can be transmitted on time and without loss. For the transmission distance, the system terminal calculates the distance based on the location coordinates of the meter device and the location coordinates of the data processing center, combined with the Euclidean distance. Meter devices with longer distances need to choose a stronger signal transmission method, or add relay nodes to expand the transmission distance. For the data type, the system terminal extracts the data stream of each dimension and matches the extraction result with the data stream template. Different types of data streams have different requirements for transmission. For example, real-time monitoring data such as current and voltage have high requirements for accuracy and transmission delay, and high priority transmission must be ensured. The transmission of some status information or user information may not have strict requirements on delay, and low priority links can be selected for transmission. Among them, the data stream template is determined based on business needs and expert decisions, and records the composition form of each type of data stream and the corresponding data sensitivity. For example, field name: current data stream, data type: floating point (current), unit: mA, A, Ah, data sensitivity: 0.3. For data sensitivity, the system terminal determines the sensitivity by matching the data flow template according to the determined data type. For data with high sensitivity, additional encryption measures or a more secure transmission link need to be taken, while for data with low sensitivity, a link with a lower security level can be selected. Afterwards, according to the obtained multidimensional data communication demand parameter set, the system terminal will perform communication diversion on the multidimensional data flow, that is, the operating data flow is classified according to the data type of the communication demand parameter of the operating data flow of each dimension. In each category, the demand similarity distance of the operating data flow of each dimension is calculated by Euclidean distance, and these demand similarity distances are compared with the preset similarity distances, and the data that meets the preset similarity distances are divided into the corresponding data flow templates. Suppose that the operation data streams of similar distances are aggregated to obtain multiple communication multi-dimensional data streams, and the required similarity distance of the operation data streams in each communication multi-dimensional data stream is less than or equal to the preset similarity distance. By arranging these multiple communication multi-dimensional data streams into multi-level communication multi-dimensional data streams, the foundation is laid for the subsequent communication link allocation, so that the system terminal can efficiently use the communication resources, avoid network congestion, and ensure that each operation data stream can be transmitted in the best way; after completing the data stream diversion, the wireless communication module will obtain a data wireless communication link list according to the network environment where the meter device is located. The list contains all available communication links, which may be different communication protocols (such as Wi-Fi , Zigbee, LoRa), frequency band (such as 2.4GHz, 868MHz, 900MHz) or transmission mode (such as serial communication, optical fiber communication, radio communication), the link list will be sorted according to different network requirements (such as transmission rate, maximum transmission distance, supported data type, security level, etc.) to ensure that the most appropriate link is selected for data transmission. The system terminal calculates the average communication rate, transmission distance, and data sensitivity of all running data flows at each level in the hierarchical communication multi-dimensional data flow to obtain the average communication demand of the level; then, based on the data type of the level, the data wireless communication link that supports this type of data transmission is extracted from the data wireless communication link list. The extracted data wireless communication links are matched with the average communication demand, and the data wireless communication links that meet the communication tolerance are obtained as the data wireless communication links of the multi-dimensional data flow of the communication at this level, that is, the data wireless communication links whose difference between the network demand supported by the data wireless communication link and the average communication demand is within the communication tolerance range, so as to avoid the waste of data wireless communication link performance; finally, the obtained data wireless communication links of each level are summarized to form multi-level data wireless communication links, which process different data flows in different levels to ensure the smooth transmission of data, minimize transmission delay and avoid data loss, and at the same time ensure the efficient operation of the entire IoT meter monitoring system. .

[0021] A data standardization program is obtained according to the data processing center, and the multidimensional operation data flow is standardized based on the data standardization program to obtain a standard multidimensional operation data flow.

[0022] In one embodiment, in the data processing center, the system terminal first obtains the data standardization program built into the data processing center, which defines how to process and convert these multi-dimensional operation data streams. Specifically, these multi-dimensional data streams include multiple different types of operation data such as current, voltage, power, temperature, meter status, user information, etc. Each data type may have different units, dimensions and precision. Therefore, the first step of the data standardization program is to identify data wild values ​​for each data, which includes identifying any abnormal values ​​that exceed the normal range (such as the peak value of current and voltage). The second step is to perform data cleaning to remove these wild value data to ensure that the data for subsequent analysis is more accurate and reliable. The third step is to perform data standardization conversion on the operation data stream of each dimension to ensure that the data between different sources and devices can be compatible and form a unified format. After this series of processing, the standard multi-dimensional operation data stream obtained is unified in terms of data format, unit, precision, etc., which not only improves the accuracy of the data, but also ensures that the data between various types of meter devices are consistent, so that the system terminal can effectively perform subsequent abnormal analysis.

[0023] Furthermore, obtaining the standard multi-dimensional operation data stream includes: According to the data standardization procedure, data wild value identification, data cleaning and data standardization conversion are obtained; the meter data application standard is obtained, and the multidimensional operation data stream is subjected to data wild value identification according to the meter data application standard to obtain the meter wild value data set; the meter wild value data set is cleaned and filtered to obtain a usable multidimensional operation data stream; based on the meter data application standard, the usable multidimensional operation data stream is subjected to data standardization conversion to obtain the standard multidimensional operation data stream.

[0024] Preferably, when the multidimensional operation data stream is transmitted to the data processing center, the system terminal will activate the data standardization program in the data processing center. This data standardization program includes three subroutines, namely, a data outlier identification subroutine, a data cleaning subroutine, and a data standardization conversion subroutine. Among them, the data outlier identification subroutine is used to identify and extract abnormal data points or values ​​beyond the normal range in the multidimensional data stream, the data cleaning subroutine is used to correct, remove or interpolate the identified abnormal data to eliminate errors, redundancy or missing parts in the data, and the data standardization conversion subroutine is used to uniformly convert the processed data into a standard format and unit to ensure the consistency and reliability of the data. Comparison and suitability for subsequent analysis; for the received multi-dimensional operation data stream, the system terminal first uses the data wild value identification subroutine to identify the data wild value of the multi-dimensional operation data stream. During the identification process, the received multi-dimensional data stream will be analyzed based on the pre-set meter data application standard (determined based on business needs and expert decisions) to identify the wild values. Wild values ​​refer to those data points that significantly deviate from the normal range or data missing at a time point, such as voltage values ​​that exceed the normal voltage range in the meter data application standard, and data missing at a predetermined time point. After wild value identification, all qualified abnormal data are extracted and form a meter wild value data set. This Some data may be caused by measurement errors, meter failures or external interference; afterwards, the identified meter wild value data set is cleaned and filtered, including removing invalid data, correcting obviously erroneous data, interpolating missing values, etc. The cleaned meter wild value data set will be synchronized to the multidimensional operation data stream to replace the original meter wild value data, thereby obtaining a usable multidimensional operation data stream, which does not contain obvious abnormal values ​​and conforms to the actual operation specifications of the meter equipment; based on the cleaned and filtered usable multidimensional operation data stream, the system terminal performs standardized conversion according to the preset meter data application standard. The purpose of standardization is to unify the format and range of the data so that data from different sources can be used together. The data can be compared and merged. If there are different units in the available multi-dimensional operation data stream (such as the current unit is milliampere and ampere), the system terminal will convert it to a standard unit (such as ampere). For data with large range differences (such as voltage may be between 0-240V, and current may be between 0-10A), the maximum and minimum value normalization operation can be performed to make the data within a relatively uniform range for subsequent processing. After standardization, the final standard multi-dimensional operation data stream has consistent units and ranges, ensuring that the multi-dimensional operation data stream from the meter equipment has high quality and high consistency, providing a reliable data basis for subsequent abnormal analysis.

[0025] Furthermore, obtaining the available multi-dimensional operation data stream includes: According to the meter wild value data set, inconsistent meter data, missing meter data and out-of-range meter data are classified; the inconsistent meter data are compared and corrected to obtain meter correction data, and the out-of-range meter data are eliminated to obtain meter normalized data; the missing meter data are interpolated and analyzed to obtain meter integrity data; based on the meter correction data, the meter normalized data and the meter integrity data, the available multi-dimensional operation data stream is obtained.

[0026] Optionally, when performing cleaning and filtering, the system terminal first classifies all data in the meter wild value data set, specifically including inconsistent meter data, missing meter data and out-of-range meter data, where inconsistent meter data is usually data that is inconsistent with the expected data pattern due to sensor failure, communication error or other technical problems, for example, the current direction conflicts with the device state (the current is negative at a certain moment, but the device is in power supply mode), non-zero power occurs at zero load, etc. Missing meter data indicates data that has not been successfully collected in certain sampling cycles, which may be due to equipment failure, signal loss or temporary power outage, etc. Out-of-range meter data refers to data that exceeds the normal working range of the meter, for example, the voltage or current value exceeds the maximum measurable range of the meter device; for inconsistent meter data, the system terminal will compare and correct it by comparing the data of other sensors. This process verifies the logic and consistency of the data and corrects abnormal data that does not conform to physical or electrical laws. , so as to obtain the meter correction data. For example, the power value should always be the product of the current and the voltage. Therefore, if the current does not match the status of the equipment, the ratio of the power value to the voltage can be used as the correction value; for the out-of-range meter data, the system terminal will delete the out-of-range data from the data set, and the data after elimination will be used as the meter normalization data. These data are in line with the normal working range of the meter and will not interfere with the subsequent analysis; for the missing meter data, the system terminal will interpolate or use other interpolation methods based on the existing valid data to fill in these missing values. The interpolation analysis can use adjacent data for interpolation. The data after interpolation will be used as the meter integrity data. These data are no longer missing and are supplemented and improved in quality; finally, the system terminal merges the meter correction data, the meter normalization data and the meter integrity data to generate a usable multi-dimensional operation data stream. This data stream contains the corrected, normalized and interpolated meter data, which provides a reliable basis for subsequent abnormal analysis and early warning.

[0027] Furthermore, obtaining the electric meter integrity data includes: Determine the missing type of the missing meter data to obtain the data missing type, and determine the sliding window length according to the data missing type and the data sampling frequency; collect neighboring data for the missing meter data based on the sliding window length to obtain a neighboring area data set; perform fitting analysis based on the neighboring area data set to construct a neighboring data regression model; perform numerical prediction and supplement on the missing meter data based on the neighboring data regression model to obtain the meter integrity data.

[0028] Optionally, the system terminal will determine the missing type of the missing meter data. The missing data may be due to various reasons, such as equipment failure, communication interruption, sensor damage, etc. In order to correctly handle these missing data, the system terminal needs to determine the type of each missing value. For example, if the data loss occurs continuously, it may be caused by sensor power failure or equipment failure. If it is intermittent loss, it may be due to signal loss or transmission problems. The system terminal will analyze the missing meter data according to the location of the missing data to obtain the data missing type involved in the missing meter data, such as continuous missing, intermittent missing, etc. After determining the missing type After that, the system terminal will determine the sliding window length in combination with the sampling frequency of the meter data. The sampling frequency refers to the time interval for data collection, usually in seconds, minutes or hours. The system terminal will select an appropriate sliding window length based on the data missing type and sampling frequency. The sliding window is used to extract data in the adjacent time period to supplement the missing data. Generally, the length of the sliding window should be sufficient to cover the changing trend of the data, but not too long to cause low processing efficiency. For continuous missing data, the system terminal will multiply the duration of the missing data (10 minutes) by the sampling frequency to obtain the sliding window length under continuous missing data. For intermittent missing data, such as the sampling frequency of current data once a minute, the sliding window length will be set to 3 to 5 minutes. This is because intermittent missing data is usually caused by temporary loss of sensor signal or short-term transmission problems. Therefore, it can be supplemented by data in a shorter time range to restore data continuity without covering data for too long a time period. Subsequently, based on the determined sliding window length, the system terminal will collect adjacent data for the missing meter data to obtain a data set of the adjacent area. After that, a fitting analysis is performed based on the obtained data set of the adjacent area. By The system terminal inputs the neighboring area data set into the linear regression model with the time as the independent variable and the meter data as the dependent variable. Through multiple fittings, the slope and intercept of the linear regression model are calculated to construct the neighboring data regression model. Then, the system terminal inputs the time corresponding to the missing meter data into the neighboring data regression model for numerical prediction supplement. The neighboring data regression model will generate corresponding prediction data according to the fitted linear regression function. The system terminal will add these prediction data to the meter integrity data to reflect the operating status of the meter equipment during the missing data period, and use them together with other data for subsequent analysis and decision support.

[0029] An electric meter abnormality analysis adaptive network is constructed by using a data mining algorithm, and abnormality analysis is performed on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter.

[0030] In one embodiment, in the data processing stage, the system terminal first establishes an electric meter abnormality indicator associated operation data set based on historical multi-dimensional operation data, and this associated operation data set includes historical multi-dimensional operation data and historical electric meter abnormality analysis indicators; then, based on the electric meter abnormality indicator associated operation data set, the system terminal combines the data characteristics of this data set with the data mining algorithm to match a model algorithm (such as a neural network, a decision tree, etc.) suitable for the current business needs, and uses the electric meter abnormality indicator associated operation data set as training data to learn and train the matched model algorithm to construct an electric meter abnormality analysis adaptive network, which is an automated integrated model that can intelligently learn normal patterns and potential anomalies in the electric meter data based on the input of a large amount of historical data and real-time data; once the electric meter abnormality is constructed Analyze the adaptive network. The system terminal will input the standard multi-dimensional operation data stream from the meter device into the network for real-time analysis. The meter abnormality analysis adaptive network will use the internal integrated model to analyze the standard multi-dimensional operation data stream one by one to generate multiple initial meter abnormal operation parameters. After that, the output layer of the meter abnormality analysis adaptive network will count the multiple initial meter abnormal operation parameters to determine the number of occurrences of each abnormal operation parameter (such as voltage, current, etc.). When the number of occurrences of a certain abnormal operation parameter exceeds the abnormal warning value, the abnormal operation parameter will be added to the meter abnormal operation parameters. After all abnormal operation parameters are counted, the meter abnormal operation parameters will be used as the final output to trigger subsequent early warning mechanisms, such as notifying maintenance personnel to conduct inspections, etc., so as to ensure the normal operation of the meter equipment.

[0031] Furthermore, the method of constructing an adaptive network for analyzing electric meter anomalies by using a data mining algorithm includes: Acquire the target of anomaly analysis of the electric meter, perform indicator splitting on the target of anomaly analysis of the electric meter, and obtain the indicator set of anomaly analysis of the electric meter; collect and acquire the historical multidimensional operation data set of the electric meter, associate and divert the historical multidimensional operation data set of the electric meter according to the indicator set of anomaly analysis of the electric meter, and obtain the operation data set associated with the abnormal indicators of the electric meter; use the data mining algorithm to select the algorithm model and perform integrated training fusion on the operation data set associated with the abnormal indicators of the electric meter, and construct an adaptive network for anomaly analysis of the electric meter.

[0032] Preferably, when constructing an adaptive network for abnormal analysis of electric meters, the system terminal first defines the abnormal analysis targets of the electric meters to be analyzed, which targets are usually the operating status or performance indicators of the equipment, such as power anomaly, voltage fluctuation, overload, overheating, etc. Once the analysis target is clarified, the target will be decomposed into several specific indicators, namely, an abnormal analysis indicator set of the electric meter, which includes but is not limited to voltage indicators (such as voltage fluctuation range, etc.), current indicators (such as current fluctuation range), power indicators (such as power fluctuation range, etc.), and equipment temperature indicators (such as temperature fluctuation range). Subsequently, the system terminal collects a large number of historical multidimensional operation data sets from the electric meter equipment, and associates these data with the defined electric meter abnormal analysis indicator set, that is, the system terminal matches the operation data of each dimension of each historical multidimensional operation data with the corresponding abnormal analysis indicator, and identifies whether the operation data of each dimension is within the fluctuation range of the corresponding indicator. After the identification is completed, the system terminal will divert the historical multidimensional operation data set, and divide the historical multidimensional operation data set into a normal historical multidimensional operation data set and an abnormal historical multidimensional operation data set. For the abnormal historical multidimensional operation data set, the system terminal can divide the historical multidimensional operation data set into a normal historical multidimensional operation data set and an abnormal historical multidimensional operation data set. The system terminal uses the corresponding abnormal parameters to identify the running data set and constructs the meter abnormal indicator associated running data set. This meter abnormal indicator associated running data set provides input data for the training of subsequent models. After obtaining the meter abnormal indicator associated running data set, the system terminal will select a suitable data mining algorithm model to process and analyze these data. Commonly used data mining algorithm models include neural networks (such as deep neural networks, convolutional neural networks), classification algorithms (such as decision trees, random forests), regression algorithms (such as linear regression, support vector regression), etc. The system terminal trains the selected data mining algorithm model through the meter abnormal indicator associated running data set, constructs multiple abnormal indicator analysis sub-models, and integrates these abnormal indicator analysis sub-models, and connects them to the input layer and output layer of the meter abnormal analysis adaptive network to construct the final meter abnormal analysis adaptive network. This network can automatically perform abnormal analysis based on the input real-time meter data and detect the abnormal operation mode of the meter. Because this network integrates multiple models, it can eliminate the limitations of a single model, increase the robustness of abnormal detection, and ensure the accuracy of abnormal analysis.

[0033] Furthermore, the construction of an adaptive network for analyzing abnormalities of electric meters includes: The abnormal analysis requirement information and data characteristic information of each indicator operation data in the abnormal indicator associated operation data set of the electric meter are respectively obtained; a data mining algorithm model set is obtained according to a data mining algorithm, and the abnormal analysis requirement information and data characteristic information are respectively matched with the data mining algorithm model set to determine the abnormal indicator matching algorithm model set; the abnormal indicator matching algorithm model set is used to perform integrated training and fusion on the abnormal indicator associated operation data set of the electric meter to obtain an electric meter abnormal analysis adaptive network.

[0034] Optionally, each abnormal analysis indicator of the electric meter device (such as voltage, current, power, etc.) will have specific analysis requirement information. For example, the current abnormality may be overload or too low, and the voltage abnormality may be excessive fluctuation or beyond the working range. The system terminal can obtain the abnormal analysis requirement information of each indicator operation data by summarizing these analysis requirement information. For each indicator operation data, it is also necessary to understand the characteristic information of the data, that is, the distribution law and timing characteristics of the data. These characteristic information can help determine the appropriate model; then, based on the abnormal analysis requirement information and data characteristic information obtained previously, the system terminal selects a suitable data mining algorithm model from the data mining algorithm model set for each abnormal analysis target, such as decision tree, random forest, support vector machine, deep neural network, ARIMA, LSTM, etc. According to the requirement information and data characteristics of each analysis indicator, the system terminal will match a suitable predetermined number of data mining algorithm models. For example, some current data may change periodically and is suitable for use. Use time series analysis models (such as ARIMA, LSTM, etc.), while some sudden power overload data are suitable for classification models (such as decision trees, random forests, etc.). The goal of the matching process is to select an effective model for each abnormal analysis target to ensure that the model can process the characteristics of the data and efficiently identify potential abnormal situations. After the selection and matching are completed, the system terminal summarizes the data mining algorithm models matched by each abnormal analysis indicator to form an abnormal indicator matching algorithm model set to ensure that the abnormal detection and analysis of each data stream can be supported by a suitable model; then, the system terminal classifies the meter abnormal indicator associated operation data set, divides a meter abnormal indicator associated operation data subset for each abnormality, and constructs multiple abnormal indicator analysis sub-models by inputting the meter abnormal indicator associated operation data subset into the corresponding model for supervised training. These abnormal indicator analysis sub-models will be combined into a meter abnormal analysis adaptive network through parallel integration, laying the foundation for accurate abnormality identification.

[0035] Furthermore, the step of obtaining an adaptive network for analyzing abnormality of an electric meter includes: The abnormal indicator matching algorithm model set is used to perform supervised training on the meter abnormal indicator associated operation data set to obtain the abnormal indicator analysis sub-model set; the abnormal indicator analysis sub-model set is integrated, fused, evaluated and optimized according to the model integration strategy to obtain the meter abnormal analysis adaptive network.

[0036] Optionally, after obtaining the abnormal indicator matching algorithm model set, the system terminal inputs the meter abnormal indicator associated operation data subset divided based on the meter abnormal indicator associated operation data set into the corresponding abnormal indicator matching algorithm model for supervised training. For example, for the meter abnormal indicator associated operation data subset that can use LSTM (such as current, voltage, etc.), the system terminal will first use LSTM to build a meter abnormality analysis model structure, including an input layer, an LSTM layer, an output layer, etc., and then input the meter abnormal indicator associated operation data subset into this model, and train it through forward propagation, loss calculation, back propagation, parameter optimization and other steps until the maximum number of iterations is reached. For the decision making model that can be used The system terminal will use the decision tree algorithm to build a meter anomaly analysis model, take the meter anomaly indicator associated operation data subset as input, use the splitting and node selection mechanism of the decision tree, generate the structure of the decision tree through the meter anomaly indicator associated operation data subset, and then use the meter anomaly indicator associated operation data subset for model training, select the best split feature by calculating the information gain or Gini coefficient, and optimize the parameters of the decision tree by back propagation until the depth of the tree or the number of nodes reaches the set threshold; after the training is completed, the system terminal will obtain a series of trained sub-models, each of which corresponds to an anomaly detection task (such as detecting current Overload, monitoring voltage fluctuations, etc.), these sub-models are added to the abnormal indicator analysis sub-model set; after obtaining the abnormal indicator analysis sub-model set, the system terminal will connect these sub-models in parallel, and then connect the parallel sub-models with the pre-set network input layer and network output layer to build an initial meter abnormality analysis adaptive network. The network output layer has a model integration strategy. This model integration strategy is based on the voting method. The output results of each sub-model will be subjected to the aforementioned voting statistics. The final output result is determined by the majority of models. In this way, the advantages of each sub-model can be used to make up for the shortcomings of a single model; after integration, the system terminal will use the data not used for training to evaluate and Tuning, the evaluation process usually uses cross-validation, accuracy, recall, F1 score and other indicators to measure the performance of the model. The system terminal optimizes the integrated model according to the evaluation results. Possible tuning methods include adjusting weights, adjusting hyperparameters (such as tree depth, SVM C parameter, etc.); Finally, the initial meter anomaly analysis adaptive network after integration and tuning will be output as the meter anomaly analysis adaptive network. This network is integrated by multiple sub-models and can process multi-dimensional data from meter equipment and perform real-time anomaly detection. The network has the ability of self-learning and optimization, and can continuously adjust and optimize its internal structure with the input of new data, thereby improving the accuracy and flexibility of anomaly detection.

[0037] A hierarchical warning mechanism is constructed, and the hierarchical warning mechanism is triggered based on the abnormal operation parameters of the electric meter to perform strategy analysis, determine the equipment warning strategy, and perform operation warning on the target electric meter equipment through the equipment warning strategy.

[0038] In one embodiment, the purpose of constructing a hierarchical warning mechanism is to set different warning levels according to the operating status and abnormal conditions of the electric meter equipment, so that when the equipment is abnormal, corresponding measures can be taken according to the severity of the abnormality. The setting of the warning mechanism usually includes multiple warning levels, each level represents a different degree of abnormal severity. Usually, the warning levels are divided into level one warning, level two warning and level three warning. The level one warning is that the deviation ratio of the abnormal operating parameters of the electric meter and the corresponding normal operating parameter range (obtained by calculating the difference between the abnormal operating parameters of the electric meter and the upper or lower limit of the normal operating parameter range, and then calculating the ratio of this difference to the upper or lower limit) is within the level one warning range, for example, slight fluctuations in current and voltage, the equipment can still operate normally, but attention is needed; the level two warning is that the deviation ratio of the abnormal operating parameters of the electric meter and the corresponding normal operating parameter range is within the level two warning range, for example, the voltage fluctuation is large, the normal operation of the equipment may be affected, although it has not yet reached a fault state; the level three warning is that the deviation ratio of the abnormal operating parameters of the electric meter and the corresponding normal operating parameter range is within the level three warning range, for example, the electric meter is overloaded, etc., which may cause equipment failure When the deviation ratio of the abnormal operation parameters of the meter reaches the set range, the system terminal will trigger the hierarchical warning mechanism and conduct a strategy analysis. The strategy of abnormal analysis will determine the measures to be taken according to the severity of the meter data. For example, for the third-level warning, the system terminal may require the meter equipment to shut down immediately and conduct an emergency inspection. For the second-level warning, an inspection reminder may be issued. For the first-level warning, it may only recommend increasing the monitoring frequency of the equipment or regular inspection. According to different abnormal parameters and abnormal levels, the equipment warning strategy will determine the most appropriate operation mode according to the current status of the equipment and actual needs to ensure that the equipment can resume normal operation in the shortest time and avoid equipment failure or damage. Through the equipment warning strategy, the system terminal will be able to issue an operation warning to the target meter equipment and monitor the operating status of the meter equipment in real time. Once the abnormal parameters trigger the warning mechanism, it will respond in real time according to the preset strategy, and notify relevant personnel through the visual interface or by sending emails, text messages, etc., so as to take timely countermeasures, thereby maximizing the operational safety and stability of the meter equipment and ensuring timely response to any potential equipment problems.

[0039] In summary, the embodiments of the present application have at least the following technical effects: First, an Internet of Things device is obtained, wherein the Internet of Things device integrates an electric meter sensor group and a wireless communication module, and the multi-dimensional operation data stream of the target electric meter device is monitored in real time through the electric meter sensor group; then, the multi-dimensional operation data stream is subjected to data transmission and analysis through the wireless communication module to obtain a multi-level data wireless communication link, and the multi-dimensional operation data stream is transmitted to a data processing center based on the multi-level data wireless communication link; further, a data standardization program is obtained according to the data processing center, and the multi-dimensional operation data stream is subjected to standardization processing based on the data standardization program to obtain a standard multi-dimensional operation data stream; then, an electric meter abnormality analysis adaptive network is constructed using a data mining algorithm, and abnormality analysis is performed on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter; finally, a hierarchical warning mechanism is constructed, and the hierarchical warning mechanism is triggered based on the abnormal operation parameters of the electric meter to perform strategy analysis, determine the equipment warning strategy, and perform operation warning on the target electric meter device through the equipment warning strategy. The technical problem of being unable to timely and accurately monitor the multi-dimensional data stream and abnormal status of the electric meter equipment due to insufficient transmission and anomaly detection mechanism of the electric meter equipment data is solved. By integrating the electric meter sensor group and the wireless communication module, and performing anomaly analysis of the data stream based on the electric meter anomaly analysis adaptive network, the technical effect of improving the data transmission efficiency of the electric meter equipment, the accuracy of anomaly analysis and the speed of early warning response is achieved.

[0040] Embodiment 2 is based on the same inventive concept as the device data analysis and early warning method based on the Internet of Things in the above embodiment. Figure 2 As shown, the present application provides an IoT-based device data analysis and early warning system, wherein the system includes: a data monitoring unit 11: obtaining an IoT device, wherein the IoT device integrates an electric meter sensor group and a wireless communication module, and real-time monitoring and obtaining a multi-dimensional operation data stream of a target electric meter device through the electric meter sensor group; a data transmission unit 12: performing data transmission and analysis on the multi-dimensional operation data stream through the wireless communication module, obtaining a multi-level data wireless communication link, and transmitting the multi-dimensional operation data stream to a data processing center based on the multi-level data wireless communication link; a standardization processing unit 13: obtaining a data standardization program according to the data processing center, and performing standardization processing on the multi-dimensional operation data stream based on the data standardization program to obtain a standard multi-dimensional operation data stream; an abnormality analysis unit 14: constructing an electric meter abnormality analysis adaptive network using a data mining algorithm, and performing abnormality analysis on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter; an operation early warning unit 15: constructing a hierarchical early warning mechanism, triggering the hierarchical early warning mechanism based on the abnormal operation parameters of the electric meter to perform strategy analysis, determine the equipment early warning strategy, and perform operation early warning on the target electric meter device through the equipment early warning strategy.

[0041] Furthermore, the data transmission unit 12 is used to perform the following method: Obtain data communication demand factors, wherein the data communication demand factors include communication rate, transmission distance, data type and data sensitivity; analyze the multidimensional operation data flow based on the data communication demand factors to obtain a multidimensional data communication demand parameter set; perform communication diversion on the multidimensional operation data flow according to the multidimensional data communication demand parameter set to obtain a hierarchical communication multidimensional data flow; obtain a data wireless communication link list through the wireless communication module, and perform communication matching with the data wireless communication link list based on the hierarchical communication multidimensional data flow to obtain the multi-level data wireless communication link.

[0042] Furthermore, the standardization processing unit 13 is used to perform the following method: According to the data standardization procedure, data wild value identification, data cleaning and data standardization conversion are obtained; the meter data application standard is obtained, and the multidimensional operation data stream is subjected to data wild value identification according to the meter data application standard to obtain the meter wild value data set; the meter wild value data set is cleaned and filtered to obtain a usable multidimensional operation data stream; based on the meter data application standard, the usable multidimensional operation data stream is subjected to data standardization conversion to obtain the standard multidimensional operation data stream.

[0043] Furthermore, the standardization processing unit 13 is used to perform the following method: According to the meter wild value data set, inconsistent meter data, missing meter data and out-of-range meter data are classified; the inconsistent meter data are compared and corrected to obtain meter correction data, and the out-of-range meter data are eliminated to obtain meter normalized data; the missing meter data are interpolated and analyzed to obtain meter integrity data; based on the meter correction data, the meter normalized data and the meter integrity data, the available multi-dimensional operation data stream is obtained.

[0044] Furthermore, the standardization processing unit 13 is used to perform the following method: The missing meter data is determined to have a missing type, and the data missing type is obtained. The sliding window length is determined according to the data missing type and the data sampling frequency. The neighboring data of the missing meter data is collected based on the sliding window length to obtain a neighboring area data set. A fitting analysis is performed based on the neighboring area data set to construct a neighboring data regression model. The missing meter data is supplemented with numerical predictions based on the neighboring data regression model to obtain the meter integrity data.

[0045] Furthermore, the abnormality analysis unit 14 is used to perform the following method: Acquire the target of anomaly analysis of the electric meter, perform indicator splitting on the target of anomaly analysis of the electric meter, and obtain the indicator set of anomaly analysis of the electric meter; collect and acquire the historical multidimensional operation data set of the electric meter, associate and divert the historical multidimensional operation data set of the electric meter according to the indicator set of anomaly analysis of the electric meter, and obtain the operation data set associated with the abnormal indicators of the electric meter; use the data mining algorithm to select the algorithm model and perform integrated training fusion on the operation data set associated with the abnormal indicators of the electric meter, and construct an adaptive network for anomaly analysis of the electric meter.

[0046] Furthermore, the abnormality analysis unit 14 is used to perform the following method: The abnormal analysis requirement information and data characteristic information of each indicator operation data in the abnormal indicator associated operation data set of the electric meter are respectively obtained; a data mining algorithm model set is obtained according to a data mining algorithm, and the abnormal analysis requirement information and data characteristic information are respectively matched with the data mining algorithm model set to determine the abnormal indicator matching algorithm model set; the abnormal indicator matching algorithm model set is used to perform integrated training and fusion on the abnormal indicator associated operation data set of the electric meter to obtain an electric meter abnormal analysis adaptive network.

[0047] Furthermore, the abnormality analysis unit 14 is used to perform the following method: The abnormal indicator matching algorithm model set is used to perform supervised training on the meter abnormal indicator associated operation data set to obtain the abnormal indicator analysis sub-model set; the abnormal indicator analysis sub-model set is integrated, fused, evaluated and optimized according to the model integration strategy to obtain the meter abnormal analysis adaptive network.

[0048] Embodiment 3, based on the same inventive concept as the device data analysis and early warning method based on the Internet of Things in the aforementioned embodiment, the present application provides a storage medium, in which a computer program is stored, and when the processor executes the computer program, the following steps are implemented: obtaining an Internet of Things device, the Internet of Things device integrates an electric meter sensor group and a wireless communication module, and real-time monitoring and obtaining a multi-dimensional operation data stream of a target electric meter device through the electric meter sensor group; performing data transmission analysis on the multi-dimensional operation data stream through the wireless communication module to obtain a multi-level data wireless communication link, and transmitting the multi-dimensional operation data stream to a data processing center based on the multi-level data wireless communication link; obtaining a data standardization program according to the data processing center, and performing standardization processing on the multi-dimensional operation data stream based on the data standardization program to obtain a standard multi-dimensional operation data stream; constructing an electric meter abnormality analysis adaptive network using a data mining algorithm, and performing abnormal analysis on the standard multi-dimensional operation data stream based on the electric meter abnormality analysis adaptive network to obtain abnormal operation parameters of the electric meter; constructing a hierarchical early warning mechanism, triggering the hierarchical early warning mechanism based on the abnormal operation parameters of the electric meter to perform strategy analysis, determine the equipment early warning strategy, and perform operation early warning on the target electric meter device through the equipment early warning strategy.

[0049] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0050] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0051] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. The device data analysis and early warning method based on the Internet of Things is characterized by: The method comprises: Acquire an Internet of Things device, wherein the Internet of Things device integrates an electric meter sensor group and a wireless communication module, and acquires a multi-dimensional operation data stream of a target electric meter device through real-time monitoring of the electric meter sensor group; Performing data transmission analysis on the multi-dimensional operation data stream through the wireless communication module to obtain a multi-level data wireless communication link, and transmitting the multi-dimensional operation data stream to a data processing center based on the multi-level data wireless communication link; Acquiring a data standardization program according to the data processing center, and performing standardization processing on the multidimensional operation data flow based on the data standardization program to obtain a standard multidimensional operation data flow; Using a data mining algorithm to construct an electric meter abnormality analysis adaptive network, based on the electric meter abnormality analysis adaptive network, perform abnormality analysis on the standard multi-dimensional operation data stream to obtain abnormal operation parameters of the electric meter; Constructing a hierarchical warning mechanism, triggering the hierarchical warning mechanism to perform strategy analysis based on the abnormal operation parameters of the electric meter, determining the equipment warning strategy, and performing operation warning on the target electric meter equipment through the equipment warning strategy; Wherein, obtaining a multi-level data wireless communication link includes: Obtaining data communication demand factors, wherein the data communication demand factors include communication rate, transmission distance, data type, and data sensitivity; Analyzing the multidimensional operation data flow based on the data communication demand factor to obtain a multidimensional data communication demand parameter set; Perform communication splitting on the multi-dimensional operation data flow according to the multi-dimensional data communication requirement parameter set to obtain a hierarchical communication multi-dimensional data flow; The wireless communication module is used to obtain a data wireless communication link list, and communication matching is performed with the data wireless communication link list based on the hierarchical communication multi-dimensional data stream to obtain the multi-level data wireless communication link.

2. The device data analysis and early warning method based on the Internet of Things as claimed in claim 1 is characterized in that: The obtaining of the standard multi-dimensional operation data stream includes: According to the data standardization procedure, data outlier identification, data cleaning and data standardization conversion are obtained; Acquire an electric meter data application standard, perform data wild value identification on the multi-dimensional operation data stream according to the electric meter data application standard, and obtain an electric meter wild value data set; Cleaning and filtering the meter wild value data set to obtain a usable multi-dimensional operation data stream; The available multi-dimensional operation data stream is converted into data standardization based on the electric meter data application standard to obtain the standard multi-dimensional operation data stream.

3. The device data analysis and early warning method based on the Internet of Things as claimed in claim 2 is characterized in that: The obtaining of available multi-dimensional operation data streams comprises: According to the meter wild value data set, inconsistent meter data, missing meter data and out-of-range meter data are classified; Comparing and correcting the inconsistent meter data to obtain meter correction data, and eliminating the out-of-range meter data to obtain meter normalization data; Performing interpolation analysis on the missing electric meter data to obtain electric meter integrity data; The available multi-dimensional operation data stream is obtained based on the meter calibration data, the meter normalization data and the meter integrity data.

4. The device data analysis and early warning method based on the Internet of Things as claimed in claim 3 is characterized in that: The obtaining of the electric meter integrity data comprises: Determine the missing type of the missing electric meter data to obtain the data missing type, and determine the sliding window length according to the data missing type and the data sampling frequency; Based on the sliding window length, adjacent data collection is performed on the missing electric meter data to obtain an adjacent area data set; Perform fitting analysis based on the adjacent area data set to construct an adjacent data regression model; The missing electricity meter data is supplemented by numerical prediction based on the neighboring data regression model to obtain the electricity meter integrity data.

5. The device data analysis and early warning method based on the Internet of Things as claimed in claim 1 is characterized in that: The method of using a data mining algorithm to construct an adaptive network for analyzing abnormalities of electric meters includes: Acquire an abnormality analysis target for an electric meter, perform index splitting on the abnormality analysis target for the electric meter, and obtain an abnormality analysis index set for the electric meter; Acquire a historical multi-dimensional operation data set of the electric meter, associate and shunt the historical multi-dimensional operation data set of the electric meter according to the electric meter abnormality analysis indicator set, and obtain an electric meter abnormality indicator associated operation data set; The data mining algorithm is used to select the algorithm model and perform integrated training fusion on the data set associated with the abnormal indicators of the electric meter, and an adaptive network for electric meter abnormality analysis is constructed.

6. The device data analysis and early warning method based on the Internet of Things as claimed in claim 5 is characterized in that: The method of constructing an adaptive network for analyzing abnormality of electric meters comprises: Respectively obtain abnormal analysis requirement information and data characteristic information of each indicator operation data in the abnormal indicator associated operation data set of the electric meter; Acquire a data mining algorithm model set according to a data mining algorithm, match the data mining algorithm model set with the abnormality analysis requirement information and the data characteristic information respectively, and determine the abnormality indicator matching algorithm model set; The abnormal indicator matching algorithm model set is used to perform integrated training and fusion on the meter abnormal indicator associated operation data set to obtain an electric meter abnormality analysis adaptive network.

7. The device data analysis and early warning method based on the Internet of Things as claimed in claim 6 is characterized in that: The method of obtaining an adaptive network for analyzing abnormality of an electric meter comprises: Using the abnormal indicator matching algorithm model set to perform supervised training on the meter abnormal indicator associated operation data set to obtain an abnormal indicator analysis sub-model set; According to the model integration strategy, the abnormal indicator analysis sub-model set is integrated, fused, evaluated and optimized to obtain the meter abnormality analysis adaptive network.

8. The device data analysis and early warning system based on the Internet of Things is characterized by: The system is used to implement the device data analysis and early warning method based on the Internet of Things according to any one of claims 1 to 7, and comprises: Data monitoring unit: obtaining an Internet of Things device, wherein the Internet of Things device integrates an electric meter sensor group and a wireless communication module, and real-time monitoring and obtaining a multi-dimensional operation data stream of a target electric meter device through the electric meter sensor group; Data transmission unit: performing data transmission analysis on the multi-dimensional operation data stream through the wireless communication module to obtain a multi-level data wireless communication link, and transmitting the multi-dimensional operation data stream to a data processing center based on the multi-level data wireless communication link; Standardization processing unit: obtaining a data standardization program according to the data processing center, and performing standardization processing on the multi-dimensional operation data flow based on the data standardization program to obtain a standard multi-dimensional operation data flow; Abnormal analysis unit: constructing an abnormal analysis adaptive network for electric meter using a data mining algorithm, performing abnormal analysis on the standard multi-dimensional operation data stream based on the abnormal analysis adaptive network for electric meter, and obtaining abnormal operation parameters of the electric meter; Operation warning unit: construct a hierarchical warning mechanism, trigger the hierarchical warning mechanism to perform strategy analysis based on the abnormal operation parameters of the meter, determine the equipment warning strategy, and perform operation warning on the target meter equipment through the equipment warning strategy.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the device data analysis and early warning method based on the Internet of Things as described in any one of claims 1 to 7 is implemented.

Images