Ground disaster hidden danger real-time detection method for electric power infrastructure

Abstract

The invention relates to a real-time detection method for a ground disaster hidden danger of an electric power infrastructure. The method comprises the following steps: S1, acquiring different types of monitoring data of a ground disaster monitoring area; independently constructing initial time sequence data of a univariate time sequence for any kind of monitoring data; s2, constructing a residual sequence according to a difference value between each data element and the average value of the front window; s3, solving two model parameters; s4, calculating latest upper and lower limit residual error thresholds by applying a second theoretical model of the extreme value theory so as to determine a normal value range; s5, calculating a residual value of the latest data; and S6, carrying out abnormity determination on the residual value. The data mode of each monitoring type is analyzed through unsupervised learning, the detection model is adjusted in real time with the self-adaptive capacity, the latest upper limit threshold value of the data can be adjusted along with the data fluctuation trend, and it is ensured that the method can adapt to various types of monitoring data.

Description

Technical Field

[0001] The invention relates to the technical field of real-time monitoring of geological disaster hazards, and in particular to a real-time detection method for geological disaster hazards of power infrastructure. Background Art

[0002] In the Yunnan, Guizhou and Sichuan regions of my country, there are abundant power generation resources, especially clean energy such as hydropower and wind power, which are widely distributed; therefore, there are a large number of power infrastructure (such as converter stations, booster stations, etc.); however, due to the complex terrain characteristics of the region, power facilities are often located in mountainous or hilly areas and are facing the threat of geological disasters for a long time, such as slope collapse, landslides, mudslides, etc. These geological disasters will not only damage power equipment, but also cause the paralysis of the power system, resulting in large-scale power outages and economic losses; therefore, how to quickly and effectively monitor and warn of geological disaster risks in power infrastructure has become one of the key issues in power system operation and maintenance.

[0003] Traditional geological disaster monitoring methods mainly rely on fixed sensors and manual inspections, which have problems such as limited monitoring range and slow response speed, and are unable to capture early signals of sudden disasters in a timely manner; in addition, there are many geological disaster risk factors such as slope collapse, landslide, and mudslide. Although different types of sensors can be deployed according to different monitoring objects, it is still a major challenge to effectively identify potential risks from a variety of monitoring data in real time; in recent years, machine learning and deep learning technologies have gradually been applied to the detection and early warning of geological disaster risks, and abnormal data beyond the normal range can be discovered; although such methods can improve detection accuracy through data-driven, they rely on a large amount of labeled data for training, and the model needs to be readjusted and trained in different application scenarios, resulting in the system's adaptability and response speed being limited, especially in complex geological environments such as Yunnan, Guizhou and Sichuan that require rapid response, and they show certain limitations in real-time response geological disaster risk monitoring. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a real-time detection method for geological disaster risks of power infrastructure, get rid of the dependence on labeled data, analyze the data pattern of each monitoring type through unsupervised learning, and have the adaptive ability to adjust the detection model in real time. It can adjust the latest upper limit threshold of the data according to the data fluctuation trend to ensure that it can adapt to various types of monitoring data.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] A method for real-time detection of geological disaster hazards of power infrastructure includes the following steps:

[0007] S1. Construct initial time series data to obtain different types of monitoring data in the geological disaster monitoring area; construct initial time series data for any type of monitoring data;

[0008] S2. Construct a residual sequence, set an initial window according to the sequence size of the initial time series data, and set the length L of the initial window Mi to 1 / 5 of the length of the initial sequence; the initial window Mi slides backward on the initial sequence, and the data elements in the initial sequence start from the position L+1. Construct a residual sequence S according to the difference between each data element Xi and the average value of the previous adjacent window Mi, and determine the upper upper threshold and the lower lower threshold of the low position by sorting the residual sequence S. The elements in the residual sequence S that exceed the upper threshold constitute an upper difference over-limit set, and the elements in the residual sequence that exceed the lower threshold constitute a lower residual over-limit set. Each element Yi in the over-limit set is the difference between the data element Xi in the initial sequence and the average value of the initial window Mi;

[0009] S3, fitting the upper difference excess set and the lower limit residual excess set to the generalized Pareto distribution GPD model, and solving two model parameters respectively;

[0010] S4. According to the model parameters and the initial time series data, the second theoretical model of extreme value theory is applied to calculate the latest upper residual threshold and lower residual threshold to determine the normal value range;

[0011] S5. Collect new data and add them into the time series. Smoothing operation is performed on the latest data element Xi, that is, the exponential moving average of the observations in the past period of time is calculated by the exponential moving average method to remove the influence of normal data fluctuations. The difference between the latest data element Xi and the weighted average constitutes the latest residual value.

[0012] S6. Perform abnormal judgment on the residual value. If it exceeds the upper or lower residual threshold, the latest data is judged to deviate from the normal range and is regarded as abnormal data, and a warning message is output; compare the residual value with the initial upper residual threshold, lower residual threshold and the latest upper residual threshold and lower residual threshold to decide whether to update the latest upper residual threshold and lower residual threshold.

[0013] In step S1, the monitoring data includes time series data of rainfall, impact force, concrete pressure, soil displacement and deep displacement.

[0014] In step S3, the upper difference excess set and the lower limit residual excess set are fitted to the generalized Pareto distribution GPD model, and the model parameters are solved by the composite parameter estimation method, as follows:

[0015] The probability weighted moment method (PWM) is used to preliminarily calculate the distribution model parameter values, and then the moment estimation method (MOM) is used to calculate the model distribution parameter values. The average of the two parameter values ​​is calculated to obtain the latest distribution model parameter values.

[0016] In step S3, respectively solving two model parameters includes the following steps:

[0017] S3.1. According to the probability weighted moment principle, the k-order probability weighted moment of the GPD model is expressed as follows:

[0018]

[0019] Formula (1), w0 is the first-order probability weighted moment; w1 is the second-order probability weighted moment; N t is the length of the excess set; E(Y) is the variance of the excess set; Y(i) is the i-th data after the time series data is sorted;

[0020] S3.2. Based on the principle of estimating the population from the sample, the GPD model parameters are calculated using the PWM method. The formula is as follows:

[0021]

[0022] S3.3. Calculate the GPD model parameters using the MOM method. The formula is as follows:

[0023]

[0024] In formula (3), μ is the mean of the excess set; S 2 is the excess set variance;

[0025] S3.4, the σ and γ parameters calculated by the composite parameter estimation method are as follows:

[0026]

[0027] In formula (IV), and It is the estimated value of the parameter to be solved in the GPD model.

[0028] In step S4, the latest upper residual threshold and lower residual threshold are calculated using the following formula:

[0029]

[0030] In formula (5), N t is the length of the excess set; q is the empirical value of the risk parameter used to determine the anomaly, that is, the probability of anomaly occurrence; n is the length of the current univariate time series.

[0031] In step S5, the exponential moving average of the observed values ​​over the past period of time is calculated, and the formula is as follows:

[0032] EMA t =α×y t +(1-α)×EMA t-1 (six)

[0033] In formula (6), EMAt is the latest exponential moving average to be calculated, yt is the observation value at the tth time point, EMAt-1 is the EMA value at the previous time point, and α is the smoothing factor, which represents the weight of the current observation value in the weighted average. It is set according to the characteristics of data fluctuations. The EMA value of the first element yt in the time series data is equal to itself.

[0034] The beneficial effects of the present invention are:

[0035] 1. No need to rely on labeled data. Unsupervised learning is used to analyze the data pattern of each monitoring type. No advance training is required. The system has the adaptive ability to adjust the detection model in real time and can adjust the latest upper limit threshold of the data according to the data fluctuation trend to ensure that it can adapt to various types of monitoring data.

[0036] 2. There is no complicated calculation in the detection method, so that each detection device has the ability of edge real-time detection, can quickly and effectively discover abnormal data in geological disaster monitoring data, and flexibly adjust the relevant parameters of the detection method according to actual conditions, further ensuring the practicality and adaptability of the system, and improving the accuracy and efficiency of real-time detection of geological disaster hazards of power infrastructure, so as to ensure the stable and safe operation of power infrastructure.

[0037] 3. Access and maintain each single variable time series data through the device channel, so that different types of data are directly isolated from each other, reducing the complexity of data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0039] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0040] Implementation 1

[0041] like Figure 1 As shown, a real-time detection method for geological disaster hazards of power infrastructure includes the following steps:

[0042] S1. Construct initial time series data to obtain different types of monitoring data in the geological disaster monitoring area; construct initial time series data for any type of monitoring data;

[0043] S2. Construct a residual sequence, set an initial window according to the sequence size of the initial time series data, and set the length L of the initial window Mi to 1 / 5 of the length of the initial sequence; the initial window Mi slides backward on the initial sequence, and the data elements in the initial sequence start from the position L+1. Construct a residual sequence S according to the difference between each data element Xi and the average value of the previous adjacent window Mi, and determine the upper upper threshold and the lower lower threshold of the low position by sorting the residual sequence S. The elements in the residual sequence S that exceed the upper threshold constitute an upper difference over-limit set, and the elements in the residual sequence that exceed the lower threshold constitute a lower residual over-limit set. Each element Yi in the over-limit set is the difference between the data element Xi in the initial sequence and the average value of the initial window Mi;

[0044] The initial time series data is a univariate time series;

[0045] The sorting of the residual sequence S determines the upper limit threshold and the lower limit threshold according to the empirical quantile (i.e. the proportion of abnormal data determined according to prior knowledge);

[0046] S3, fitting the upper difference excess set and the lower limit residual excess set to the generalized Pareto distribution GPD model, and solving two model parameters respectively;

[0047] S4. According to the model parameters and the initial time series data, the second theoretical model of extreme value theory is applied to calculate the latest upper residual threshold and lower residual threshold to determine the normal value range;

[0048] S5. Collect new data and add them into the time series. Smoothing operation is performed on the latest data element Xi, that is, the exponential moving average of the observations in the past period of time is calculated by the exponential moving average method to remove the influence of normal data fluctuations. The difference between the latest data element Xi and the weighted average constitutes the latest residual value.

[0049] S6. Perform abnormal judgment on the residual value. If it exceeds the upper or lower residual threshold, the latest data is judged to deviate from the normal range and is regarded as abnormal data, and a warning message is output; compare the residual value with the initial upper residual threshold, lower residual threshold and the latest upper residual threshold and lower residual threshold to decide whether to update the latest upper residual threshold and lower residual threshold.

[0050] In step S1, the monitoring data includes time series data of rainfall, impact force, concrete pressure, soil displacement and deep displacement.

[0051] In step S3, the upper difference excess set and the lower limit residual excess set are fitted to the generalized Pareto distribution GPD model, and the model parameters are solved by the composite parameter estimation method, as follows:

[0052] The probability weighted moment method (PWM) is used to preliminarily calculate the distribution model parameter values, and then the extremely fast moment estimation method (MOM) is used to calculate the model distribution parameter values. The two parameter values ​​are averaged to obtain the latest distribution model parameter values.

[0053] In step S3, respectively solving two model parameters includes the following steps:

[0054] S3.1. According to the probability weighted moment principle, the k-order probability weighted moment of the GPD model is expressed as follows:

[0055]

[0056] Formula (1), w0 is the first-order probability weighted moment; w1 is the second-order probability weighted moment; N t is the length of the excess set; E(Y) is the variance of the excess set; Y(i) is the i-th data after the time series data is sorted;

[0057] S3.2. Based on the principle of estimating the population from the sample, the GPD model parameters are calculated using the PWM method. The formula is as follows:

[0058]

[0059] S3.3. Calculate the GPD model parameters using the MOM method. The formula is as follows:

[0060]

[0061] In formula (3), μ is the mean of the excess set; S 2 is the excess set variance;

[0062] S3.4, the σ and γ parameters calculated by the composite parameter estimation method are as follows:

[0063]

[0064] In formula (IV), and It is the estimated value of the parameter to be solved in the GPD model.

[0065] In step S4, the latest upper residual threshold and lower residual threshold are calculated using the following formula:

[0066]

[0067] In formula (5), N t is the length of the excess set; q is the empirical value of the risk parameter used to determine the anomaly, that is, the probability of anomaly occurrence; n is the length of the current univariate time series.

[0068] This method can be used to calculate the upper and lower residual thresholds to determine the normal value range.

[0069] In step S5, the exponential moving average of the observed values ​​over the past period of time is calculated, and the formula is as follows:

[0070] EMA t =α×y t +(1-α)×EMA t-1 (six)

[0071] In formula (6), EMAt is the latest exponential moving average to be calculated, yt is the observation value at the tth time point, EMAt-1 is the EMA value at the previous time point, and α is the smoothing factor, which represents the weight of the current observation value in the weighted average. It is set according to the characteristics of data fluctuations. The EMA value of the first element yt in the time series data is equal to itself.

[0072] In step S6, the abnormality judgment does not directly use the latest data element Xi, but uses the residual value Xi' calculated in step S5 for judgment. If Xi' is greater than the latest upper residual threshold, or less than the lower residual threshold, it is considered that the latest data Xi exceeds the normal range and is judged as abnormal data; if the latest residual value Xi' is between the initial upper residual threshold and the latest upper residual threshold, Xi' is added to the upper limit excess set; similarly, if Xi' is between the initial lower residual threshold and the latest lower residual threshold, it is added to the lower limit excess set; when a new element is added to the excess set, the latest residual threshold can be updated through steps S3 and S4.

[0073] A real-time detection device for geological disaster hazards of power infrastructure, comprising:

[0074] The data acquisition module is configured with multiple data access channels, each of which can be connected to a sensor terminal to collect a type of detection data.

[0075] The data conversion module is equipped with a processor and a memory, and is used to convert the analog signals collected by the sensor terminal into numerical signals. Each channel can be configured with independent conversion rules to convert and store the collected raw data. Each channel maintains a time series, that is, the data collected by each channel constitutes a univariate time series data.

[0076] The analysis and processing module is configured with a memory and a processor. The processor runs a detection method to detect the monitoring data collected by each channel in real time, and uses the single variable time series data composed of each type of data as the processing object. When new data arrives, it detects in real time whether the data is abnormal.

[0077] The configuration module is used to receive and store the original conversion rules from the system configuration and the relevant parameter values ​​required by the real-time detection method.

[0078] A real-time detection system for geological disaster hazards of power infrastructure, the system comprising:

[0079] The data storage module is used to receive and store output data from a real-time detection device for geological disaster hazards of multiple power infrastructures.

[0080] The data display module is used to display the alarm status, the raw data curves of each sensor, and the upper and lower thresholds.

[0081] The system configuration module is used to configure the calculation rules for converting sensor analog signals into numerical signals, allowing users to adjust parameters such as empirical abnormality probability and smoothing factor according to actual detection needs.

[0082] The output module outputs the collected raw data, the upper and lower residual thresholds dynamically updated during the anomaly determination process, the exponential moving average and other process data and anomaly detection results.

[0083] There is no need to rely on labeled data. The data pattern of each monitoring type is analyzed through unsupervised learning. No advance training is required, and the detection model has the adaptive ability to adjust the latest upper limit threshold of the data in real time. It can adjust the data according to the data fluctuation trend to ensure that it can adapt to various types of monitoring data.

[0084] There is no complicated calculation in the detection method, so each detection device has the ability of edge real-time detection, which can quickly and effectively discover abnormal data in geological disaster monitoring data, and flexibly adjust the relevant parameters of the detection method according to actual conditions, further ensuring the practicality and adaptability of the system, and improving the accuracy and efficiency of real-time detection of geological disaster hazards of power infrastructure, so as to ensure the stable and safe operation of power infrastructure.

[0085] Each single variable time series data is accessed and maintained through the device channel, so that different types of data are directly isolated from each other, reducing the complexity of data processing.

[0086] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A real-time detection method for geological disaster hazards of power infrastructure, characterized by: The following steps are involved: S1. Construct initial time series data to obtain different types of monitoring data in the geological disaster monitoring area; construct initial time series data for any type of monitoring data; S2. Construct a residual sequence, set an initial window according to the sequence size of the initial time series data, and set the length L of the initial window Mi to 1 / 5 of the length of the initial sequence; the initial window Mi slides backward on the initial sequence, and the data elements in the initial sequence start from the position L+1. Construct a residual sequence S according to the difference between each data element Xi and the average value of the previous adjacent window Mi, and determine the upper upper threshold and the lower lower threshold of the low position by sorting the residual sequence S. The elements in the residual sequence S that exceed the upper threshold constitute an upper difference over-limit set, and the elements in the residual sequence that exceed the lower threshold constitute a lower residual over-limit set. Each element Yi in the over-limit set is the difference between the data element Xi in the initial sequence and the average value of the initial window Mi; S3, fitting the upper difference excess set and the lower limit residual excess set to the generalized Pareto distribution GPD model, and solving two model parameters respectively; S4. According to the model parameters and the initial time series data, the second theoretical model of extreme value theory is applied to calculate the latest upper residual threshold and lower residual threshold to determine the normal value range; S5. Collect new data and add them into the time series. Smoothing operation is performed on the latest data element Xi, that is, the exponential moving average of the observations in the past period of time is calculated by the exponential moving average method to remove the influence of normal data fluctuations. The difference between the latest data element Xi and the weighted average constitutes the latest residual value. S6. Perform abnormality determination on the residual value. If the residual value exceeds the upper or lower residual threshold, the latest data is considered to be abnormal data if it deviates from the normal range, and an early warning message is output; The residual value is compared with the initial upper residual threshold, the lower residual threshold and the latest upper residual threshold and the lower residual threshold to determine whether to update the latest upper residual threshold and the lower residual threshold.

2. A method for real-time detection of geological disaster hazards of power infrastructure according to claim 1, characterized in that: In step S1, the monitoring data includes time series data of rainfall, impact force, concrete pressure, soil displacement and deep displacement.

3. A method for real-time detection of geological disaster hazards of power infrastructure according to claim 1, characterized in that: In step S3, the upper difference excess set and the lower limit residual excess set are fitted to the generalized Pareto distribution GPD model, and the model parameters are solved by the composite parameter estimation method, as follows: The probability weighted moment method (PWM) is used to preliminarily calculate the distribution model parameter values, and then the moment estimation method (MOM) is used to calculate the model distribution parameter values. The average of the two parameter values ​​is calculated to obtain the latest distribution model parameter values.

4. A method for real-time detection of geological disaster hazards of power infrastructure according to claim 1, characterized in that: In step S3, respectively solving two model parameters includes the following steps: S3.

1. According to the probability weighted moment principle, the k-order probability weighted moment of the GPD model is expressed as follows: Formula (1), w0 is the first-order probability weighted moment; w1 is the second-order probability weighted moment; N t is the length of the excess set; E(Y) is the variance of the excess set; Y(i) is the i-th data after the time series data is sorted; S3.

2. Based on the principle of estimating the population from the sample, the GPD model parameters are calculated using the PWM method. The formula is as follows: S3.

3. Calculate the GPD model parameters using the MOM method. The formula is as follows: In formula (3), μ is the mean of the excess set; S 2 is the excess set variance; S3.4, the σ and γ parameters calculated by the composite parameter estimation method are as follows: In formula (IV), and It is the estimated value of the parameter to be solved in the GPD model.

5. A method for real-time detection of geological disaster hazards of power infrastructure according to claim 1, characterized in that: In step S4, the latest upper residual threshold and lower residual threshold are calculated using the following formula: In formula (5), N t is the length of the excess set; q is the empirical value of the risk parameter used to determine the anomaly, that is, the probability of anomaly occurrence; n is the length of the current univariate time series.

6. A method for real-time detection of geological disaster hazards of power infrastructure according to claim 1, characterized in that: In step S5, the exponential moving average of the observed values ​​over the past period of time is calculated, and the formula is as follows: EMA t = α × y t + (1 - α) × EMA t-1 (Six) In formula (6), EMAt is the latest exponential moving average to be calculated, yt is the observation value at the tth time point, EMAt-1 is the EMA value at the previous time point, and α is the smoothing factor, which represents the weight of the current observation value in the weighted average. It is set according to the characteristics of data fluctuations. The EMA value of the first element yt in the time series data is equal to itself.

Images