Intelligent switch abnormal signal detection method based on power line carrier communication

Abstract

The invention provides an intelligent switch abnormal signal detection method based on power line carrier communication, and the method comprises the steps: 1, synchronous collection and dynamic preprocessing of intelligent switch signals, 2, multi-dimensional feature fusion extraction, 3, abnormal signal classification and recognition, 4, real-time detection and warning, and 5, adaptive anti-interference optimization. Step 6, data remote transmission; step 7, dynamic clustering mining of an abnormal power consumption mode; the multi-dimensional feature fusion is combined with the power utilization big data time sequence feature of the intelligent switch and the time frequency feature extracted by wavelet transform, so that the transient change and the abnormal feature of the signal can be captured, and the accuracy of abnormal signal detection is improved; the power utilization big data of the intelligent switch can be processed in real time through synchronous acquisition and dynamic preprocessing, abnormal signals are quickly classified and recognized through a logistic regression model, and alarm can be immediately triggered when the abnormal signals appear in combination with a real-time detection and warning mechanism.

Description

Technical Field

[0001] The present invention relates to a method for detecting abnormal signals of intelligent switches based on power line carrier communication, and particularly relates to the technical field of abnormal signal detection. Background Art

[0002] With the rapid development of smart grid and Internet of Things technologies, the demand for intelligent management of power systems is increasing day by day. As a key device in the power system, intelligent switches undertake important functions such as power distribution, load control, and fault protection. However, during actual operation, intelligent switches may cause unstable operation of the power system or even lead to safety accidents due to abnormal power line carrier communication signals, equipment aging, external interference, etc. Therefore, real-time detection and accurate identification of abnormal signals of intelligent switches have become an important topic for ensuring the safe and stable operation of power systems.

[0003] Traditional abnormal signal detection methods mainly rely on manual inspections and simple threshold judgments, and have problems such as low detection efficiency, high false alarm rate, and poor real-time performance. In addition, power line carrier communication signals are easily affected by factors such as environmental noise and frequency offset interference, resulting in inaccurate signal feature extraction and further reducing the reliability of abnormal detection.

[0004] Traditional methods mostly rely on single features (such as voltage and current), which are difficult to comprehensively reflect the operating state of intelligent switches; power line carrier communication signals are easily affected by environmental noise and frequency offset interference, and existing methods have insufficient research on signal preprocessing and anti-interference optimization; there is little research on dynamic clustering and in-depth mining of abnormal power consumption patterns, and it is difficult to discover potential abnormal power consumption behaviors. Summary of the Invention

[0005] Object of the Invention: To propose a method for detecting abnormal signals of intelligent switches based on power line carrier communication to solve the above problems existing in the prior art.

[0006] Technical Solution: A method for detecting abnormal signals of intelligent switches based on power line carrier communication includes:

[0007] Step 1: Synchronous acquisition and dynamic preprocessing of intelligent switch signals,

[0008] Step 2: Multi-dimensional feature fusion extraction,

[0009] Step 3: Classification and identification of abnormal signals,

[0010] Step 4: Real-time detection and warning,

[0011] Step 5: Adaptive anti-interference optimization.

[0012] In a further embodiment, step one is to collect power line carrier signals in real time through the power line carrier communication module built into the intelligent switch; preprocess the collected signals, including filtering, denoising, and normalization processing. The normalization processing is carried out through the following formula:

[0013]

[0014] In the formula, x, x q are the big data of the intelligent switch before and after normalization processing, and x max , x min are the maximum and minimum values in the big data sample of the intelligent switch; x mean is the average value of the big data sample of the intelligent switch.

[0015] In a further embodiment, step two is to analyze the preprocessed signals and extract the characteristic parameters of the signals; extract the time-frequency characteristics of the signals through wavelet transform to capture the transient changes of the signals. The specific steps are as follows: decompose the signals by wavelet to obtain wavelet coefficients at different scales; calculate the energy distribution of the wavelet coefficients and extract the time-frequency characteristics;

[0016] Utilize the time-series characteristics of the big data of the intelligent switch's electricity consumption to construct a multi-dimensional feature vector for subsequent abnormal signal recognition. The specific steps are as follows: extract the time-series characteristics of the electricity consumption data of the intelligent switch, combine the time-series characteristics with the time-frequency characteristics extracted by wavelet transform, and construct a multi-dimensional feature vector.

[0017] In a further embodiment, step three is to perform variable-frequency sampling processing on the power line carrier communication abnormal signals based on the multi-dimensional feature vector to obtain an intermediate-frequency filtered signal, and remove the frequency offset information through a numerically controlled signal oscillator; based on the extracted characteristic parameters, construct a logistic regression model to classify and identify abnormal intelligent switches. The Sigmoid function is cited in the logistic regression model, and the expression of the Sigmoid function is:

[0018]

[0019] In the formula, h θ is the big data of the intelligent switch with electricity consumption characteristics; is the multi-dimensional vector formed by the time-series characteristics of the big data of the intelligent switch; θ t is the time-series characteristic parameter of the big data of the grid intelligent switch at time t; if h θ (x) = 1, it indicates that the identified intelligent switch is abnormal, otherwise it is normal.

[0020] In a further embodiment, in step four, according to the classification result of the logistic regression model, the power consumption category features are extracted, the power consumption data of the power system are collected, the data of different categories are distinguished, divided into two categories of normal and abnormal, and the marked data are used for model training; the trained model is applied to new power consumption data to automatically determine whether it belongs to the abnormal category; when an abnormal signal is detected, the alarm mechanism is triggered, the characteristic information of the abnormal signal is recorded, and the user or the operation and maintenance personnel are notified through the communication module of the intelligent switch for processing.

[0021] In a further embodiment, in step five, the signal is decomposed by wavelet transform at multiple scales, the wavelet packet coefficients at different scales are extracted, and the signal singular points are obtained through normalization processing to form a signal filter; according to the changes in the actual operating environment, the parameters and thresholds of the model are dynamically adjusted, and the following threshold calculation function is used to dynamically adjust the abnormal detection threshold:

[0022]

[0023] In the formula, T is the threshold; σ is the standard deviation; j is the scale of signal decomposition; n is the signal length.

[0024] In a further embodiment, the power line carrier communication module converts the electric energy metering data into a carrier signal; uses the FSK modulation method to convey digital information by changing the frequency of the signal, and embeds the digital information into the carrier signal for transmission; the specific steps are as follows:

[0025] The electric energy metering data is converted into a carrier signal through an FSK modulator, and the FSK modulator selects different frequencies for transmission according to the binary value of the input data; the modulated carrier signal is transmitted through the power line to the receiving end; the power line serves as the transmission medium to carry the carrier signal; at the receiving end, the FSK demodulator restores the carrier signal to the original data, and the demodulator judges the binary value of the data according to the received frequency.

[0026] In a further embodiment, according to the transmission result, by extracting the key features in the abnormal power consumption data, the specific pattern of abnormal power consumption is identified; the abnormal data is clustered to obtain the average value of the abnormal data cluster, and the edge data is extracted to obtain the abnormal power consumption information.

[0027] Beneficial effects: The present invention proposes an abnormal signal detection method for an intelligent switch based on power line carrier communication. The multi-dimensional feature fusion combines the time-series features of the power consumption big data of the intelligent switch and the time-frequency features extracted by wavelet transform, which can capture the transient changes and abnormal features of the signal, and improve the accuracy of abnormal signal detection; the synchronous acquisition and dynamic preprocessing can process the power consumption big data of the intelligent switch in real time, and quickly classify and identify abnormal signals through a logistic regression model. Combining real-time detection and warning mechanisms, an alarm can be triggered immediately when an abnormal signal appears.

[0028] The logistic regression model and clustering analysis are used to achieve the automatic classification of abnormal signals and the intelligent mining of abnormal power consumption patterns; through the power line carrier device and FSK modulation, the remote transmission of electric energy metering data is realized; through dynamic clustering analysis, key features are extracted from the abnormal power consumption data to identify specific abnormal power consumption patterns. By analyzing the abnormal data clusters, potential abnormal power consumption behaviors are discovered, providing data support for the fault prevention and optimized management of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the overall process of the present invention.

[0030] Figure 2 It is a flowchart of the abnormal signal detection of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] To solve the problems existing in the prior art, the present invention provides an intelligent switch abnormal signal detection method based on power line carrier communication. By synchronously collecting the large amount of power consumption data of the intelligent switch and combining technologies such as dynamic preprocessing, multi-dimensional feature fusion, logistic regression classification, and adaptive anti-interference optimization, accurate detection and real-time warning of abnormal signals are achieved.

[0032] The following further specifically describes the solution through embodiments and in conjunction with the drawings.

[0033] In this application, we propose an intelligent switch abnormal signal detection method based on power line carrier communication, including the following steps:

[0034] Step 1: Synchronously collect and dynamically preprocess the intelligent switch signals, and normalize the large amount of power consumption data of the intelligent switch; real-time collect power line carrier signals (voltage, current, power parameters) through the power line carrier communication module built in the intelligent switch, and preprocess the collected signals, including filtering, denoising, and normalization processing, to eliminate environmental noise and interference. The normalization processing is carried out through the following formula:

[0035]

[0036] In the formula, x and x q are the large amount of intelligent switch data before and after normalization processing, x max , x min are the maximum and minimum values in the intelligent switch data sample; x mean is the average value of the intelligent switch data sample.

[0037] Step 2: Multi-dimensional feature fusion extraction. Use the time-series features of the big data of intelligent switch power consumption to extract signal features; analyze the preprocessed signal to extract the characteristic parameters of the signal; extract the time-frequency features of the signal through wavelet transform to capture the transient changes of the signal. The specific steps are as follows: perform wavelet decomposition on the signal to obtain wavelet coefficients at different scales; calculate the energy distribution of the wavelet coefficients and extract time-frequency features (such as power consumption change rate, power fluctuation).

[0038] Use the time-series features of the big data of intelligent switch power consumption to construct a multi-dimensional feature vector for subsequent abnormal signal recognition. The specific steps are as follows: extract the time-series features of the power consumption data of the intelligent switch (power consumption change rate, power fluctuation), and combine the time-series features with the time-frequency features extracted by wavelet transform to construct a multi-dimensional feature vector.

[0039] Step 3: Abnormal signal classification and recognition. Perform frequency conversion sampling on the abnormal signals of power line carrier communication to remove frequency offset information, and classify and recognize the abnormal intelligent switches through a logistic regression model; based on the multi-dimensional feature vector, perform frequency conversion sampling on the abnormal signals of power line carrier communication to obtain an intermediate frequency filtered signal, and remove the frequency offset information through a numerically controlled signal oscillator; based on the extracted characteristic parameters, construct a logistic regression model to classify and recognize the abnormal intelligent switches. The Sigmoid function is cited in the logistic regression model, and the expression of the Sigmoid function is:

[0040]

[0041] In the formula, h θ is the big data of intelligent switches of power consumption characteristics; is the multi-dimensional vector formed by the time-series characteristics of the big data of intelligent switches; θ t is the time-series characteristic parameter of the big data of the power grid intelligent switch at time t; h θ (x) = 1 indicates that the identified intelligent switch is abnormal, otherwise it is normal.

[0042] Step 4: Real-time detection and warning. Extract the power consumption category features and perform abnormal determination; according to the classification results of the logistic regression model, extract the power consumption category features, collect the power consumption data of the power system (real-time or historical data of voltage, current, power, power consumption), distinguish different types of data, divide them into normal and abnormal categories, and use the labeled data for model training; apply the trained model to new power consumption data to automatically determine whether it belongs to the abnormal category; when an abnormal signal is detected, trigger the alarm mechanism, record the characteristic information of the abnormal signal, and notify the user or the operation and maintenance personnel to handle it through the communication module of the intelligent switch.

[0043] Step 5: Adaptive anti-interference optimization. Extract signal features through wavelet transform to complete abnormal signal detection. Decompose the signal into multiple scales through wavelet transform, extract wavelet packet coefficients at different scales, and obtain signal singular points through normalization processing to form a signal filter. Dynamically adjust the parameters and thresholds of the model according to the changes in the actual operating environment to improve the adaptability and accuracy of the detection method. Dynamically adjust the abnormal detection threshold using the following threshold calculation function:

[0044]

[0045] In the formula, T is the threshold; σ is the standard deviation; j is the scale of signal decomposition; n is the signal length.

[0046] The power line carrier communication module converts the electric energy metering data into a carrier signal. Using the FSK modulation method, digital information is conveyed by changing the frequency of the signal, and the digital information is embedded in the carrier signal for transmission. The specific steps are as follows:

[0047] Convert the electric energy metering data into a carrier signal through an FSK modulator. The FSK modulator selects different frequencies for transmission according to the binary value (0 or 1) of the input data. Transmit the modulated carrier signal through the power line to the receiving end. The power line serves as the transmission medium to carry the carrier signal. At the receiving end, use an FSK demodulator to restore the carrier signal to the original data. The demodulator judges the binary value of the data according to the received frequency.

[0048] According to the transmission results, identify the specific patterns of abnormal electricity consumption by extracting the key features (such as abnormal increase in electricity consumption, abnormal change in electricity consumption time) in the abnormal electricity consumption data. Cluster the abnormal data to obtain the average value of the abnormal data cluster, and extract the edge data to obtain the abnormal electricity consumption information.

[0049] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as a limitation of the present invention itself. Various changes can be made in its form and details without departing from the spirit and scope of the present invention defined by the appended claims.

Claims

1. An abnormal signal detection method for an intelligent switch based on power line carrier communication, characterized in that, Including: Step 1: Synchronous acquisition and dynamic preprocessing of intelligent switch signals, Step 2: Multi-dimensional feature fusion extraction, Step 3: Abnormal signal classification and recognition, Step 4: Real-time detection and warning, Step 5: Adaptive anti-interference optimization.

2. The intelligent switch abnormal signal detection method based on power line carrier communication according to claim 1, wherein In Step 1, the power carrier signal is collected in real time through the power carrier communication module built in the intelligent switch; the collected signal is preprocessed, including filtering, denoising and normalization processing. The normalization processing is carried out through the following formula: where x, x q are the big data of intelligent switches before and after normalization processing, x max , x min are the maximum and minimum values in the big data sample of intelligent switches; x mean is the average value of the big data sample of intelligent switches.

3. The intelligent switch abnormal signal detection method based on power line carrier communication according to claim 1, characterized in that In Step 2, the preprocessed signal is analyzed to extract the characteristic parameters of the signal; the time-frequency characteristics of the signal are extracted through wavelet transform to capture the transient changes of the signal. The specific steps are as follows: the signal is wavelet decomposed to obtain wavelet coefficients at different scales; the energy distribution of the wavelet coefficients is calculated to extract the time-frequency characteristics; Using the time-series characteristics of the big data of the intelligent switch power consumption, a multi-dimensional feature vector is constructed for subsequent abnormal signal recognition; The specific steps are: extracting the time-series characteristics of the power consumption data of the intelligent switch, combining the time-series characteristics with the time-frequency characteristics extracted by wavelet transform, and constructing a multi-dimensional feature vector.

4. The intelligent switch abnormal signal detection method based on power line carrier communication according to claim 1, wherein, In Step 3, based on the multi-dimensional feature vector, the power carrier communication abnormal signal is subjected to variable-frequency sampling processing to obtain an intermediate-frequency filtered signal, and the frequency offset information is removed through a numerically controlled signal oscillator; based on the extracted characteristic parameters, a logistic regression model is constructed to classify and recognize the abnormal intelligent switch. The Sigmoid function is cited in the logistic regression model. The expression of the Sigmoid function is: where h θ is the big data of intelligent switches of electricity consumption characteristics; is the multi-dimensional vector formed by the time series characteristics of the big data of intelligent switches; θ t is the time series characteristic parameter of the big data of the grid intelligent switch at time t; h θ (x) = 1 indicates that the recognized intelligent switch is abnormal, otherwise it is normal.

5. The intelligent switch abnormal signal detection method based on power line carrier communication according to claim 1, characterized in that, In Step 4, according to the classification result of the logistic regression model, the power consumption category characteristics are extracted, the power consumption data of the power system are collected, the data of different categories are distinguished, divided into two categories: normal and abnormal, and the marked data are used for model training; the trained model is applied to the new power consumption data to automatically judge whether it belongs to the abnormal category; when an abnormal signal is detected, the alarm mechanism is triggered, the characteristic information of the abnormal signal is recorded, and the user or the operation and maintenance personnel are notified through the communication module of the intelligent switch for processing.

6. The intelligent switch abnormal signal detection method based on power line carrier communication according to claim 1, wherein In Step 5, the signal is multi-scale decomposed through wavelet transform, the wavelet packet coefficients at different scales are extracted, and the signal singular points are obtained through normalization processing to form a signal filter; according to the change of the actual operating environment, the parameters and thresholds of the model are dynamically adjusted, and the following threshold calculation function is used to dynamically adjust the abnormal detection threshold: In the formula, T is the threshold; σ is the standard deviation; j is the scale of signal decomposition; n is the signal length.

7. A method for detecting abnormal signals of an intelligent switch based on power line carrier communication according to claim 1, characterized in that The power carrier communication module converts the electric energy metering data into a carrier signal; uses the FSK modulation method to convey digital information by changing the frequency of the signal, and embeds the digital information into the carrier signal for transmission; the specific steps are as follows: The electric energy metering data is converted into a carrier signal through an FSK modulator. The FSK modulator selects different frequencies for transmission according to the binary value of the input data; the modulated carrier signal is transmitted to the receiving end through the power line; the power line is used as the transmission medium to carry the carrier signal; at the receiving end, the FSK demodulator restores the carrier signal to the original data, and the demodulator judges the binary value of the data according to the received frequency.

8. The intelligent switch abnormal signal detection method based on power line carrier communication according to claim 7, characterized in that According to the transmission results, by extracting the key features in the abnormal power consumption data, the specific patterns of abnormal power consumption are identified; the abnormal data is clustered to obtain the average value of the abnormal data cluster, and the edge data is extracted to complete the mining of abnormal power consumption information.

Images