Power Internet of Things anomaly detection method and system

Abstract

The invention belongs to the technical field of communication systems, and discloses a power internet of things anomaly detection method and system. Operating data are collected and preprocessed, an electric power data set is obtained, and the operating data comprise current data and voltage data; calculating an associated reference value between the power devices based on the power data set, and establishing a power detection model; the power data set and the associated reference value are detected, and the detected power data set and the associated reference value are classified and processed in real time, so that the abnormality of the power equipment in a certain time in the future can be effectively predicted, and the loss caused by the abnormality of the power Internet of Things is greatly reduced. The method has the remarkable advantages of high anomaly prediction capability, good auxiliary level and high risk prevention and control capability.

Description

Technical Field

[0001] The present invention relates to the technical field of power anomaly monitoring, and more specifically, to a power Internet of Things anomaly detection method and system. Background Art

[0002] The power Internet of Things is the application of the Internet of Things in the smart grid. It is the result of the development of information and communication technology to a certain stage. It will effectively integrate communication infrastructure resources and power system infrastructure resources, improve the level of informationization of the power system, improve the utilization efficiency of the existing infrastructure of the power system, and provide important technical support for power grid power generation, transmission, transformation, distribution, and power consumption. With the rapid development of society, people are paying more and more attention to the demand for electricity. Today, the stability of power supply is directly related to thousands of households. The current method of abnormal detection of the power Internet of Things is mostly to install sensors for detection. Compared with the traditional manual fixed-cycle detection, this method reduces the labor cost to a certain extent, but it is still in a passive state where the abnormality can only be obtained after the abnormality occurs. The patent application with the publication number CN117113261B discloses a power Internet of Things detection method, which stores the Internet of Things data in an independent database, converts the Internet of Things data into numerical values, and then determines multiple Internet of Things curves, and then determines the data similarity of multiple Internet of Things terminals based on the Internet of Things curves, and then determines the correlation of multiple Internet of Things terminals. After comprehensive judgment, the correlation evaluation parameter is introduced to finally improve the integrity of the recognition process; However, although the above technologies have improved the integrity of the identification process through several measures, the smart grid is a huge facility composed of many power equipment. Therefore, the abnormal detection of power equipment is particularly important. If there are safety hazards in the power equipment, it is very easy to cause the use of the smart grid to stop, and then cause great economic losses. Therefore, how to quickly identify the abnormal state in the power Internet of Things and predict the existence of abnormal state is related to the normal life guarantee of thousands of households. In view of this, the present invention proposes a power Internet of Things anomaly detection method and system to solve the above problems. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solutions: comprising: A data acquisition module is used to collect operation data and pre-process the operation data to obtain a power data set, where the operation data includes current data and voltage data; Furthermore, the methods of preprocessing the operation data include data cleaning and data conversion; Data cleaning methods include: Preset the value range, and automatically filter the data that exceeds the value range based on the SQL constraint checking tool, and mark it as abnormal value; Replace abnormal values ​​in the current data and the voltage data with the mean value of the current data or the voltage data to obtain first data; Data conversion methods include: Converting the first data from an analog signal to a digital signal based on an analog-to-digital converter to obtain a power data set; The data mining module calculates the correlation reference values ​​between power equipment based on the power data set and establishes a power detection model; Furthermore, the method of calculating the correlation reference value between the power equipment includes: Based on the preset intervals of the power equipment, the power equipment is divided into Segment, take N groups of power data sets of each power equipment, and substitute them into the calculation formula: Get the associated reference value ,in and are the detection values ​​of two power devices in the same group N, the detection value is the sum of the current value and the voltage value, C1 and V1 are the averages of the detection values ​​of the two power devices, and n is the number of detection values; Furthermore, the methods of establishing the power detection model include: Collect J groups of associated reference values ​​as sample sets, and divide the sample sets into 70%J training set, 15%J test set, and 15%J validation set; A prediction model is established based on the training set. The input layer of the prediction model is used to input the sample set, and the output layer is used to output the predicted observation value. The prediction model adopts any one of the linear regression model, random forest model and time series analysis model. The mathematical expression of the prediction model is: , and obtain the predicted observed value ,in is the observed value at a time point, is a constant term, is the autoregression coefficient, is the error term, is the next moment after the time point; Through S rounds of iterative training, the prediction model parameters are stabilized and the power detection model is obtained; A detection execution module is used to detect the power data set and the associated reference value, and classify the detected power data set and the associated reference value to obtain a processing result; Further, the method of detecting the power data set and the associated reference value includes: Identify power data sets and associated reference values, identify power data sets and associated reference values ​​based on the Z-score statistical method and remove outliers; Through cross-validation, the results of the detected power data set and the associated reference values ​​are compared. The result comparison is to compare L data items of the same data set, and the power data set and the associated reference values ​​after removing outliers are verified; Furthermore, the methods of classification and real-time processing include: Based on Fourier transform signal processing technology, voltage characteristics and current characteristics are extracted from the second data, the voltage characteristics include voltage amplitude and harmonic content, and the current characteristics include current amplitude and current harmonics; A detection threshold set is preset, and the data items in the voltage characteristics and current characteristics are detected in real time based on the direct traversal method of Python, and the voltage characteristics and current characteristics are analyzed. The detection threshold set includes the voltage amplitude threshold interval [P1, P2], the harmonic content threshold Q1, the current amplitude threshold interval [U1, U2] and the current harmonic threshold E1; Based on the Z-score statistical method, the data items in the voltage characteristics and current characteristics are marked, that is, the voltage amplitude is Wa, the harmonic content is Wb, the current amplitude is Wc, and the current harmonic is Wd; Based on the voltage amplitude threshold interval [P1, P2], when Wa is greater than P1 and less than P2, a notification SMS is sent to the staff receiving end through the communication unit; when Wa is greater than P2, a warning SMS is sent to the staff receiving end through the communication unit; Based on the harmonic content threshold Q1, when Wb is greater than Q1, a warning SMS is sent to the staff receiving end through the communication unit; Based on the current amplitude threshold interval [U1, U2], when Wc is greater than U1 and less than U2, a notification SMS is sent to the staff receiving end through the communication unit; when Wc is greater than U2, a warning SMS is sent to the staff receiving end through the communication unit; Based on the current harmonic threshold E1, when Wd is greater than E1, a warning SMS is sent to the staff receiving end through the communication unit; Preset reference threshold interval [ ,0, ], when the associated reference value is [ ,0] or [0, ], a warning SMS is sent to the staff receiving terminal through the communication unit; Based on the reference threshold interval [ ,0, ], when the pre-observation value is [ ,0] or [0, ], a warning SMS is sent to the staff receiving terminal through the communication unit; The notification SMS includes content that explains the existence of abnormal data in specific data items in the power data set and associated reference values ​​and prompts staff to log in to the system. The warning SMS includes content that explains the existence of serious abnormalities in specific data items in the power data set and associated reference values ​​and requires staff to take immediate action. The early warning SMS includes content that explains the existence of serious abnormalities in specific data items in the pre-observation values ​​and requires staff to take immediate action.

[0004] The technical effects and advantages of the present invention are as follows: The present invention realizes the detection of anomalies that may occur in power equipment during operation based on the correlation between power equipment through calculation and analysis of associated reference values. Compared with the traditional power equipment anomaly detection method, the present invention not only has the function of real-time detection of whether there are anomalies in power equipment, but also can execute different commands according to the level of anomaly. The established power detection model is used to further analyze the associated reference values. Based on multiple sets of past and continuous associated reference values, the trend of future associated reference values ​​is predicted, which can effectively predict the anomalies in power equipment within a certain period of time in the future, and greatly reduce the losses caused by anomalies in the power Internet of Things. Generally speaking, the present invention has the significant advantages of strong anomaly prediction ability, good auxiliary level and high risk prevention and control ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 It is a schematic diagram of the power Internet of Things anomaly detection system of the present invention; Figure 2 It is a schematic diagram of the power Internet of Things anomaly detection method of the present invention. DETAILED DESCRIPTION

[0006] 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 only 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.

[0007] Example 1 See also Figure 1 As shown, the power Internet of Things anomaly detection method and system described in this embodiment include: A data acquisition module is used to collect operation data and pre-process the operation data to obtain a power data set, where the operation data includes current data and voltage data; Furthermore, the methods of preprocessing the operation data include data cleaning and data conversion; Data cleaning methods include: Preset the value range, and automatically filter the data that exceeds the value range based on the SQL constraint checking tool, and mark it as abnormal value; It should be explained that the numerical range is the reasonable range of current data and voltage data. If there is no data reading suddenly in a set of voltage data, it is an abnormal value beyond the reasonable range; Replace abnormal values ​​in the current data and the voltage data with the mean value of the current data or the voltage data to obtain first data; Data conversion methods include: Converting the first data from an analog signal to a digital signal based on an analog-to-digital converter to obtain a power data set; The data mining module calculates the correlation reference values ​​between power equipment based on the power data set and establishes a power detection model; Furthermore, the method of calculating the correlation reference value between the power equipment includes: Based on the preset intervals of the power equipment, the power equipment is divided into Segment, take N groups of power data sets of each power equipment, and substitute them into the calculation formula: Get the associated reference value ,in and are the detection values ​​of two power devices in the same group N, the detection value is the sum of the current value and the voltage value, C1 and V1 are the averages of the detection values ​​of the two power devices, and n is the number of detection values; It should be explained that the preset interval of the power equipment is based on the interval length or the number of power equipment, and the correlation reference value reflects the value of detecting the failure of the power equipment based on the correlation between two power equipments in the same group; Further, the methods of establishing the power detection model include: Collect J groups of associated reference values ​​as sample sets, and divide the sample sets into 70%J training set, 15%J test set, and 15%J validation set; A prediction model is established based on the training set. The input layer of the prediction model is used to input the sample set, and the output layer is used to output the predicted observation value. The prediction model adopts any one of the linear regression model, random forest model and time series analysis model. It should be explained that the prediction model is a type of data prediction model, which is a technical model that predicts future data trends based on past historical data; The mathematical expression of the prediction model is: , and obtain the predicted observed value ,in is the observed value at a time point, is a constant term, is the autoregression coefficient, is the error term, is the next moment after the time point; It needs to be explained that express right The impact of; assumption -U(0, ), that is, it has a mean of 0 and a variance of of normal distribution, and for different , are independent of each other, the error term is the random variation value that the model cannot explain, and the predicted observation value at the next moment The observed value , a constant and a random error value jointly determined; Through S rounds of iterative training, the prediction model parameters are stabilized and the power detection model is obtained; A detection execution module is used to detect the power data set and the associated reference value, and classify the detected power data set and the associated reference value to obtain a processing result; Further, the method of detecting the power data set and the associated reference value includes: Identify power data sets and associated reference values, identify power data sets and associated reference values ​​based on the Z-score statistical method and remove outliers; Through cross-validation, the results of the detected power data set and the associated reference values ​​are compared. The result comparison is to compare L data items of the same data set, and the power data set and the associated reference values ​​after removing outliers are verified; It should be explained that the result comparison is to compare L data items under the same set of data sets. When the comparison results are the same, whether the data items meet the preset rules is used to evaluate the accuracy of removing outliers; Further, the methods for classification and real-time processing include: Based on Fourier transform signal processing technology, voltage characteristics and current characteristics are extracted from the second data, the voltage characteristics include voltage amplitude and harmonic content, and the current characteristics include current amplitude and current harmonics; A detection threshold set is preset, and the data items in the voltage characteristics and current characteristics are detected in real time based on the direct traversal method of Python, and the voltage characteristics and current characteristics are analyzed. The detection threshold set includes the voltage amplitude threshold interval [P1, P2], the harmonic content threshold Q1, the current amplitude threshold interval [U1, U2] and the current harmonic threshold E1; Based on the Z-score statistical method, the data items in the voltage characteristics and current characteristics are marked, that is, the voltage amplitude is Wa, the harmonic content is Wb, the current amplitude is Wc, and the current harmonic is Wd; Based on the voltage amplitude threshold interval [P1, P2], when Wa is greater than P1 and less than P2, a notification SMS is sent to the staff receiving end through the communication unit; when Wa is greater than P2, a warning SMS is sent to the staff receiving end through the communication unit; Based on the harmonic content threshold Q1, when Wb is greater than Q1, a warning SMS is sent to the staff receiving end through the communication unit; Based on the current amplitude threshold interval [U1, U2], when Wc is greater than U1 and less than U2, a notification SMS is sent to the staff receiving end through the communication unit; when Wc is greater than U2, a warning SMS is sent to the staff receiving end through the communication unit; Based on the current harmonic threshold E1, when Wd is greater than E1, a warning SMS is sent to the staff receiving end through the communication unit; Preset reference threshold interval [ ,0, ], when the associated reference value is [ ,0] or [0, ], a warning SMS is sent to the staff receiving terminal through the communication unit; Based on the reference threshold interval [ ,0, ], when the pre-observation value is [ ,0] or [0, ], a warning SMS is sent to the staff receiving end through the communication unit; The notification SMS includes the content of explaining that there are abnormal data in specific data items in the power data set and the associated reference values ​​and prompting the staff to log in to the system; the warning SMS includes the content of explaining that there are serious abnormalities in specific data items in the power data set and the associated reference values ​​and requiring the staff to take immediate action; the early warning SMS includes the content of explaining that there are serious abnormalities in specific data items in the pre-observation values ​​and requiring the staff to take immediate action; In this embodiment, the beneficial effect is that through the calculation and analysis of the associated reference values, it is realized that the abnormalities that may occur in the operation of the power equipment are detected based on the correlation between the power equipment. Compared with the traditional power equipment abnormality detection method, the present invention not only has the function of real-time detection of whether there are abnormalities in the power equipment, but also can execute different commands according to the level of the abnormality. The established power detection model is used to further analyze the associated reference values. Based on multiple sets of past and continuous associated reference values, the trend of future associated reference values ​​is predicted, which can effectively predict the abnormalities in the power equipment within a certain period of time in the future, and greatly reduce the losses caused by the abnormalities in the power Internet of Things. Generally speaking, the present invention has the significant advantages of strong abnormality prediction ability, good auxiliary level and high risk prevention and control ability.

[0008] Example 2 See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, and a method and system for detecting abnormalities in an electric power Internet of Things are provided, including: S1: collecting operation data, and preprocessing the operation data to obtain an electric power data set, where the operation data includes current data and voltage data; S2: Calculate the correlation reference values ​​between power equipment based on the power data set and establish a power detection model; S3: Detecting the power data set and the associated reference values, classifying the detected power data set and the associated reference values, and processing them in real time.

[0009] Example 3 This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the above-mentioned power Internet of Things anomaly detection method is implemented.

[0010] Since the electronic device introduced in this embodiment is an electronic device used to implement a method for detecting anomalies in an electric power Internet of Things in the embodiment of this application, based on the method for detecting anomalies in an electric power Internet of Things introduced in the embodiment of this application, a person skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application is not described in detail here. As long as a person skilled in the art implements an electronic device used in a method for detecting anomalies in an electric power Internet of Things in the embodiment of this application, it belongs to the scope of protection of this application.

[0011] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0012] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A power Internet of Things anomaly detection system, characterized in that: include: A data acquisition module is used to collect operation data and pre-process the operation data to obtain a power data set, where the operation data includes current data and voltage data; The data mining module calculates the correlation reference values ​​between power equipment based on the power data set and establishes a power detection model; The detection execution module is used to detect the power data set and the associated reference value, and classify the detected power data set and the associated reference value to obtain a processing result.

2. The power Internet of Things anomaly detection system according to claim 1 is characterized in that: The methods of preprocessing the operational data include data cleaning and data conversion; Data cleaning methods include: Preset the value range, and automatically filter the data that exceeds the value range based on the SQL constraint checking tool, and mark it as abnormal value; Replace abnormal values ​​in the current data and the voltage data with the mean value of the current data or the voltage data to obtain first data; Data conversion methods include: The first data is converted from an analog signal to a digital signal based on an analog-to-digital converter to obtain a power data set.

3. The power Internet of Things anomaly detection system according to claim 2 is characterized in that: Methods for calculating the correlation reference values ​​between power equipment include: Based on the preset intervals of the power equipment, the power equipment is divided into Segment, take N groups of power data sets of each power equipment, and substitute them into the calculation formula: Get the associated reference value ,in and are the detection values ​​of two power devices in the same group N, the detection value is the sum of the current value and the voltage value, C1 and V1 are the means of the detection values ​​of the two power devices, and n is the number of detection values.

4. The power Internet of Things anomaly detection system according to claim 3 is characterized in that: Ways to build a power detection model include: Collect J groups of associated reference values ​​as sample sets, and divide the sample sets into 70%J training set, 15%J test set, and 15%J validation set; A prediction model is established based on the training set. The input layer of the prediction model is used to input the sample set, and the output layer is used to output the predicted observation value. The prediction model adopts any one of the linear regression model, random forest model and time series analysis model. The mathematical expression of the prediction model is: , and obtain the predicted observed value ,in is the observed value at a time point, is a constant term, is the autoregression coefficient, is the error term, is the next moment after the time point; Through S rounds of iterative training, the prediction model parameters are stabilized and the power detection model is obtained.

5. The power Internet of Things anomaly detection system according to claim 4 is characterized in that: Methods for detecting power data sets and associated reference values ​​include: Identify power data sets and associated reference values, identify power data sets and associated reference values ​​based on the Z-score statistical method and remove outliers; Through cross-validation, the results of the detected power data set and the associated reference values ​​are compared. The result comparison is a comparison of L data items in the same data set. The power data set and the associated reference values ​​after removing outliers are verified.

6. The power Internet of Things anomaly detection system according to claim 5 is characterized in that: Ways to classify and process in real time include: Based on Fourier transform signal processing technology, voltage characteristics and current characteristics are extracted from the second data, the voltage characteristics include voltage amplitude and harmonic content, and the current characteristics include current amplitude and current harmonics; A detection threshold set is preset, and the data items in the voltage characteristics and current characteristics are detected in real time based on the direct traversal method of Python, and the voltage characteristics and current characteristics are analyzed. The detection threshold set includes the voltage amplitude threshold interval [P1, P2], the harmonic content threshold Q1, the current amplitude threshold interval [U1, U2] and the current harmonic threshold E1; Based on the Z-score statistical method, the data items in the voltage characteristics and current characteristics are marked, that is, the voltage amplitude is Wa, the harmonic content is Wb, the current amplitude is Wc, and the current harmonic is Wd; Based on the voltage amplitude threshold interval [P1, P2], when Wa is greater than P1 and less than P2, a notification SMS is sent to the staff receiving end through the communication unit; when Wa is greater than P2, a warning SMS is sent to the staff receiving end through the communication unit; Based on the harmonic content threshold Q1, when Wb is greater than Q1, a warning SMS is sent to the staff receiving end through the communication unit; Based on the current amplitude threshold interval [U1, U2], when Wc is greater than U1 and less than U2, a notification SMS is sent to the staff receiving end through the communication unit; when Wc is greater than U2, a warning SMS is sent to the staff receiving end through the communication unit; Based on the current harmonic threshold E1, when Wd is greater than E1, a warning SMS is sent to the staff receiving end through the communication unit; Preset reference threshold interval [ ,0, ], when the associated reference value is [ ,0] or [0, ], a warning SMS is sent to the staff receiving terminal through the communication unit; Based on the reference threshold interval [ ,0, ], when the pre-observation value is [ ,0] or [0, ], a warning SMS is sent to the staff receiving terminal through the communication unit; The notification SMS includes content that explains the existence of abnormal data in specific data items in the power data set and associated reference values ​​and prompts staff to log in to the system. The warning SMS includes content that explains the existence of serious abnormalities in specific data items in the power data set and associated reference values ​​and requires staff to take immediate action. The early warning SMS includes content that explains the existence of serious abnormalities in specific data items in the pre-observation values ​​and requires staff to take immediate action.

7. A method for detecting anomalies in an electric power Internet of Things, which is implemented based on an anomaly detection system for an electric power Internet of Things according to any one of claims 1 to 6, characterized in that: S1: Collecting operation data and preprocessing the operation data to obtain a power data set, where the operation data includes current data and voltage data; S2: Calculate the correlation reference values ​​between power equipment based on the power data set and establish a power detection model; S3: Detecting the power data set and the associated reference values, classifying the detected power data set and the associated reference values, and processing them in real time.

Images