DTU safety monitoring method, system and device based on power consumption information and storage medium

By obtaining and processing DTU power consumption information in real time, using unsupervised encoding-decoding learning framework and intelligent early warning system, dynamically identifying and responding to DTU abnormal behavior, the problem of insufficient adaptability of traditional monitoring methods is solved, and the self-adaptation of DTU security monitoring is realized Adaptability and efficiency.

CN119071190AInactive Publication Date: 2024-12-03SHENZHEN FRIENDCOM TECH DEV

Patent Information

Application Number
CN202411226042.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-12-03
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure CN119071190A_ABST
    Figure CN119071190A_ABST
Patent Text Reader

Abstract

The invention discloses a DTU safety monitoring method, system and device based on power consumption information and a storage medium, and the method comprises the steps: obtaining the power consumption information of a DTU in real time, and carrying out the preprocessing of the power consumption information of the DTU, and obtaining a DTU power consumption sequence; respectively extracting time domain features and frequency domain features of the DTU power consumption sequence, and fusing to generate a DTU power consumption feature set; based on historical power consumption characteristics of the DTU, unsupervised coding-decoding learning is carried out on the DTU power consumption characteristic set, and based on a preset first threshold value, a normal power consumption behavior and an abnormal power consumption behavior are identified through reconstruction error analysis; on the basis of historical power consumption characteristics of the DTU, abnormal probability distribution is calculated, and according to second power consumption information of the monitored DTU, the boundary of a first threshold value is adjusted in a self-adaptive mode; and when an abnormal power consumption behavior is detected, automatically generating a countermeasure instruction and an adjustment suggestion through a preset intelligent early warning system according to preset rules and strategies. The self-adaptive, efficient, accurate and timely safety monitoring of the DTU power consumption behavior is realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to a DTU security monitoring method, system, device and storage medium based on power consumption information. Background Art

[0002] DTU (Data Transfer Unit) is a device dedicated to remote data communication. It is widely used in many industries such as electricity, water conservancy, petroleum, heat, and environmental protection. With the rapid development of smart grids, DTU has become a key link in ensuring the stable operation of power systems. However, DTU may encounter various abnormal conditions such as data overload, power supply abnormality, channel abnormality, and hardware and software failures during operation. These abnormal conditions are often first reflected in abnormal changes in power consumption behavior patterns. Traditional security monitoring methods mostly rely on hard threshold judgments or simple statistical analysis, lack dynamic adaptability, and cannot adaptively adjust monitoring strategies according to real-time monitoring conditions, which reduces the ability to respond to new abnormal conditions. The ability to identify complex and changeable abnormal behaviors is limited, and it is difficult to meet the growing security monitoring needs of smart grids. Therefore, there is an urgent need for a DTU security monitoring method that can adaptively, accurately, and timely identify DTU abnormal power consumption behavior, which can not only improve the security protection level of the power grid, but also effectively prevent systemic accidents caused by DTU failures. Summary of the invention

[0003] The present invention provides a DTU security monitoring method, system, equipment and storage medium based on power consumption information.

[0004] A first aspect of the present invention provides a DTU security monitoring method based on power consumption information, and the DTU security monitoring method based on power consumption information includes: Acquire power consumption information of the DTU in real time, and pre-process the power consumption information of the DTU to obtain a DTU power consumption sequence; Extracting time domain features and frequency domain features of the power consumption sequence of the DTU respectively, and fusing them to generate a DTU power consumption feature set; Based on the historical power consumption characteristics of the DTU, unsupervised encoding-decoding learning is performed on the DTU power consumption characteristic set, and based on a preset first threshold, normal power consumption behavior and abnormal power consumption behavior are identified through reconstruction error analysis; Based on the historical power consumption characteristics of the DTU, calculate the abnormal probability distribution, and adaptively adjust the limit of the first threshold according to the second power consumption information of the monitored DTU; When abnormal power consumption behavior is detected, the preset intelligent early warning system will automatically generate response instructions and adjustment suggestions based on preset rules and strategies.

[0005] Optionally, in a first implementation manner of the first aspect of the present invention, the acquiring the power consumption information of the DTU in real time, and preprocessing the power consumption information of the DTU to obtain a DTU power consumption sequence includes: Through current sensors and voltage sensors, current and voltage are monitored in real time, and power consumption is calculated; Converting the analog signal into a digital signal through an analog-to-digital converter to generate power consumption information of the DTU; Regularly read the power consumption information of the DTU in different states; Performing multi-level filtering on the power consumption information of the DTU to obtain smoothed power consumption information of the DTU; According to the maximum-minimum normalization principle, the power consumption information of the DTU is standardized to obtain a DTU power consumption sequence.

[0006] Optionally, in a second implementation of the first aspect of the present invention, time domain features and frequency domain features are extracted from the power consumption sequence of the DTU respectively, and a DTU power consumption feature set is generated by fusing them, including: Perform time domain feature extraction on the power consumption sequence of the DTU: The power consumption sequence is converted into a three-dimensional tensor according to [sample quantity, time step, power consumption value feature dimension] and input into the input layer; Set 32 ​​one-dimensional convolution kernels with a size of 3 time steps to slide on the time axis, and convolve to obtain local time series features; The convolution formula is: ; in, is the activation function; Weight is the convolution kernel weight; is a fragment of the input sequence; Bias is the bias term; Set ReLU as the activation function, reduce the time dimension through the maximum pooling layer, and extract key features; Setting a fully connected layer and an output layer to identify or predict a specific power consumption mode or state, outputting a time domain feature sequence of the DTU, and generating a time domain feature set of the DTU; Perform frequency domain feature extraction on the power consumption sequence of the DTU: Performing a fast Fourier transform operation on the time domain characteristic sequence of the DTU to generate a spectrum diagram; The Fourier transform formula is: ; Where X[k] represents the complex value of the transformed frequency domain signal at the kth frequency point; x[n] is the value of the time domain characteristic sequence of the DTU at the nth time point; N is the length of the time domain characteristic sequence of the DTU; It is a complex exponential function and the basic rotation factor in Fourier transform, which is used to decompose the time domain characteristic sequence of DTU into the superposition of different frequency components; Analyze the spectrum graph to identify frequency components and potential frequency patterns based on the frequency and amplitude of periodic power consumption fluctuations; The power spectrum density and frequency peak position of a specific frequency are selected to construct the frequency domain feature set of the DTU.

[0007] Optionally, in a third implementation manner of the first aspect of the present invention, time domain features and frequency domain features are extracted from the power consumption sequence of the DTU respectively, and the features are fused to generate a DTU power consumption feature set, further comprising: The time domain feature set of the DTU and the frequency domain feature set of the DTU are integrated to generate a DTU power consumption feature set, including: Time-aligning the time domain feature set of the DTU and the frequency domain feature set of the DTU; Performing a weighted fusion operation on the time domain feature set of the DTU and the frequency domain feature set of the DTU to generate the DTU power consumption feature set; The calculation expression of the weighted fusion is: ; Among them, C is a composite feature vector that integrates time domain and frequency domain information to form a DTU power consumption feature set; T and F represent the feature vectors of the time domain feature set and the frequency domain feature set, respectively; and are the weights of time domain and frequency domain features respectively.

[0008] Optionally, in a fourth implementation manner of the first aspect of the present invention, the performing unsupervised encoding-decoding learning on the DTU power consumption feature set based on the historical power consumption characteristics of the DTU, and identifying normal power consumption behavior and abnormal power consumption behavior through reconstruction error analysis, including: Based on the historical power consumption characteristics of the DTU, a stacked deep autoencoder model is constructed; Set up 2 hidden layers, each layer includes linear transformation and nonlinear activation function ReLU; Compress and encode the historical power consumption characteristics of the DTU and reconstruct them by decoding; The calculation expression of the hidden layer in the stacked deep autoencoder model is: ; Among them, the The activation value of the layer; It is the nonlinear activation function ReLU; and are the weight matrix and bias vector of the layer respectively; Based on the mean square error, optimize the weights and biases to minimize the reconstruction error; The expression of the mean square error is: ; Where L is the mean square error, which is used to measure the difference between the model output and the actual value; B is the number of samples; is the bth sample input; is the reconstructed output of the bth sample; Input the DTU power consumption feature set into the stacked deep autoencoder model, perform decoding and reconstruction, and output a first reconstruction error; When the first reconstruction error does not exceed a preset first threshold, it is identified as normal power consumption behavior; When the first reconstruction error exceeds a preset first threshold, it is identified as abnormal power consumption behavior.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, the calculating of the abnormal probability distribution based on the historical power consumption characteristics of the DTU, and adaptively adjusting the limit of the first threshold according to the second power consumption information of the monitored DTU include: Based on Gaussian distribution, characterize the historical power consumption characteristics of the DTU and calculate the abnormal probability distribution of the reconstruction error; The calculation formula of the abnormal probability distribution is: ; in, is the reconstruction error; is the abnormal probability distribution, which represents the probability density function of the error and reflects the possibility of the error occurring; and are the mean and standard deviation of historical errors, indicating the central position of the error distribution and the degree of data dispersion; Calculating the dynamic threshold according to the second power consumption information of the DTU; The calculation formula of the dynamic threshold is: ; in, is a factor that controls the strictness of the threshold, and a larger Indicates a higher standard of anomaly detection; ; and are the mean and standard deviation of historical errors, respectively, indicating the central position of the error distribution and the degree of data dispersion.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, when abnormal power consumption behavior is detected, a preset intelligent early warning system is used to automatically generate countermeasure instructions and adjustment suggestions according to preset rules and strategies, including: When abnormal power consumption behavior is detected, the intelligent early warning system immediately sends an early warning to the maintenance personnel through at least one method including SMS, email and APP push, and marks the abnormal time period and possible causes; Evaluate the stability of the communication link through indicators such as bit error rate and packet loss rate; Dynamically adjust the operating frequency and power output of the DTU based on the stability of the communication link and task requirements monitored in real time: quickly transmit data when the stability of the communication link is high, and reduce the transmission rate when the stability of the communication link is poor; Adjust the allocation of DTU data transmission tasks to avoid high-load operations during concentrated time periods and reduce abnormal peak power consumption.

[0011] The second aspect of the present invention provides a DTU safety monitoring system based on power consumption information, the DTU safety monitoring system based on power consumption information includes: an acquisition module, a processing module and an early warning module; each module is communicated with each other; Optionally, in a first implementation manner of the second aspect of the present invention, the acquisition module is used to acquire power consumption information of the DTU in real time; Optionally, in a second implementation of the second aspect of the present invention, the processing module is used to preprocess the power consumption information of the DTU to obtain a DTU power consumption sequence; extract time domain features and frequency domain features of the DTU power consumption sequence respectively, and fuse them to generate a DTU power consumption feature set; based on the historical power consumption characteristics of the DTU, perform unsupervised encoding-decoding learning on the DTU power consumption feature set, and based on a preset first threshold, identify normal power consumption behavior and abnormal power consumption behavior through reconstruction error analysis; based on the historical power consumption characteristics of the DTU, calculate the abnormal probability distribution, and adaptively adjust the limit of the first threshold according to the second power consumption information of the monitored DTU; and also be used to execute the above-mentioned DTU security monitoring method based on power consumption information; Optionally, in a third implementation of the second aspect of the present invention, the early warning module is used to automatically generate response measures instructions and adjustment suggestions through a preset intelligent early warning system according to preset rules and strategies when abnormal power consumption behavior is detected.

[0012] A third aspect of the present invention provides a DTU security monitoring device based on power consumption information, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the DTU security monitoring device based on power consumption information executes the above-mentioned DTU security monitoring method based on power consumption information.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned DTU safety monitoring method based on power consumption information.

[0014] Beneficial effects: The DTU security monitoring method based on power consumption information provided by the present invention, on the one hand, introduces an unsupervised encoding-decoding learning framework, and uses the historical power consumption characteristics of DTU to train the extracted power consumption feature set. This process can not only learn the complex patterns of normal power consumption behavior, but also automatically distinguish between normal and abnormal power consumption behaviors through reconstruction error analysis without manual labeling, and issue early warning responses and strategies in a timely manner, which greatly reduces the burden of manual intervention and improves the level of automation of monitoring. Combined with the preset first threshold, the framework can flexibly adapt to the normal power consumption range in different scenarios, reducing the probability of false alarms and missed alarms. On the other hand, through in-depth analysis of historical data, the abnormal probability distribution is calculated, and the limit of the first threshold can be dynamically adjusted according to the second power consumption information monitored in real time. This intelligent adjustment mechanism ensures the adaptability and accuracy of the monitoring system, and even when dealing with constantly changing complex working conditions and self-generated or environmental conditions, it can maintain a high degree of monitoring efficiency and effectively respond to various unexpected changes.

[0015] It realizes adaptive, efficient, accurate and timely security monitoring of DTU power consumption behavior, effectively reduces risks and significantly enhances the security protection capabilities of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of an embodiment of a DTU security monitoring method based on power consumption information in an embodiment of the present invention; Figure 2 The figure is a schematic diagram of an embodiment of a DTU security monitoring system based on power consumption information in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] Embodiments of the present invention provide a DTU security monitoring method, system, device and storage medium based on power consumption information. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of a DTU security monitoring method based on power consumption information in an embodiment of the present invention includes: 101. Acquire power consumption information of a DTU in real time, and pre-process the power consumption information of the DTU to obtain a DTU power consumption sequence; It is understandable that the execution subject of the present invention may be a DTU security monitoring system based on power consumption information, or may be a terminal or a server, which is not specifically limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0019] Specifically, the real-time acquisition of the power consumption information of the DTU and preprocessing of the power consumption information of the DTU to obtain a DTU power consumption sequence includes: Through current sensors and voltage sensors, current and voltage are monitored in real time, and power consumption is calculated; Converting the analog signal into a digital signal through an analog-to-digital converter to generate power consumption information of the DTU; Regularly read the power consumption information of the DTU in different states; Performing multi-level filtering on the power consumption information of the DTU to obtain smoothed power consumption information of the DTU; According to the maximum-minimum normalization principle, the power consumption information of the DTU is standardized to obtain a DTU power consumption sequence.

[0020] More specifically, by installing high-precision current sensors (such as Hall effect current sensors) and voltage sensors (such as resistor divider voltage sensors) at key nodes of DTU, these sensors are directly connected to the power supply line of DTU to achieve real-time and uninterrupted monitoring of current and voltage. Through a high-performance analog-to-digital converter (ADC, Analog-to-Digital Converter), its sampling rate is set to at least thousands of times per second to meet the requirements of transient response analysis of the power system to ensure the quality of the collected signals. When the analog signals of current and voltage are input to the analog-to-digital converter, they are immediately converted into high-resolution digital signals. Using the collected current and voltage digital signals, the power consumption value is calculated in real time according to the formula P=UI (i.e. rate = voltage × current); the power consumption information is read from different working states of DTU at preset time intervals (such as every minute or every second). This includes but is not limited to standby, communication, reclosing operation and other states to fully understand the energy consumption mode of DTU. In order to eliminate noise interference, a multi-stage filtering strategy is adopted for the original power consumption data. First, a low-pass filter is applied to remove high-frequency noise, then a median filter is used to further reduce the impact of random errors, and finally a sliding average filter is used to smooth the data sequence to ensure that the power consumption information obtained is more stable and reliable. The power consumption information after filtering is standardized using the maximum-minimum normalization principle, that is, the power consumption value at each time point is scaled so that all data is mapped to the [0, 1] interval. Normalization processing helps the consistency and comparison of subsequent algorithm processing.

[0021] 102. Extract time domain features and frequency domain features of the power consumption sequence of the DTU respectively, and fuse them to generate a DTU power consumption feature set; Specifically, the time domain features and frequency domain features of the power consumption sequence of the DTU are extracted respectively, and the DTU power consumption feature set is generated by fusion, including: Perform time domain feature extraction on the power consumption sequence of the DTU: The power consumption sequence is converted into a three-dimensional tensor according to [sample quantity, time step, power consumption value feature dimension] and input into the input layer; Set 32 ​​one-dimensional convolution kernels with a size of 3 time steps to slide on the time axis, and convolve to obtain local time series features; The convolution formula is: ; in, is the activation function; Weight is the convolution kernel weight; is a fragment of the input sequence; Bias is the bias term; Set ReLU as the activation function, reduce the time dimension through the maximum pooling layer, and extract key features; Setting a fully connected layer and an output layer to identify or predict a specific power consumption mode or state, outputting a time domain feature sequence of the DTU, and generating a time domain feature set of the DTU; More specifically, let's analyze the daily power consumption sequence of a DTU in a week, with a total of 7 days of data, and each piece of data contains 24 hours of power consumption values ​​(i.e., the time step is 24). First, the data is processed into a tensor of shape [7, 24, 1], where the last dimension is the feature dimension, which is only a single channel in this case, i.e., the power consumption value. 32 convolution kernels with a time step size of 3 (i.e., each kernel covers 3 hours of data), and ReLU is used as the activation function. The convolution operation scans the entire sequence to capture local time series features such as daily power consumption peaks and valleys. Subsequently, the time dimension is reduced through the maximum pooling layer to extract key features. Finally, the fully connected layer and the output layer are used to identify or predict specific power consumption modes or states. It can extract useful time series features such as periodic patterns and abnormal mutations from continuous power consumption sequences, and can also provide richer input features for anomaly detection in the stacked deep autoencoder model, improving the accuracy and robustness of detection.

[0022] Perform frequency domain feature extraction on the power consumption sequence of the DTU: Performing a fast Fourier transform operation on the time domain characteristic sequence of the DTU to generate a spectrum diagram; The Fourier transform formula is: ; Where X[k] represents the complex value of the transformed frequency domain signal at the kth frequency point; x[n] is the value of the time domain characteristic sequence of the DTU at the nth time point; N is the length of the time domain characteristic sequence of the DTU; It is a complex exponential function and the basic rotation factor in Fourier transform, which is used to decompose the time domain characteristic sequence of DTU into the superposition of different frequency components; Analyze the spectrum graph to identify frequency components and potential frequency patterns based on the frequency and amplitude of periodic power consumption fluctuations; The power spectrum density and frequency peak position of a specific frequency are selected to construct the frequency domain feature set of the DTU.

[0023] Specifically, extracting the time domain features and the frequency domain features of the power consumption sequence of the DTU respectively, and fusing them to generate a DTU power consumption feature set, further comprising: The time domain feature set of the DTU and the frequency domain feature set of the DTU are integrated to generate a DTU power consumption feature set, including: Time-aligning the time domain feature set of the DTU and the frequency domain feature set of the DTU; Performing a weighted fusion operation on the time domain feature set of the DTU and the frequency domain feature set of the DTU to generate the DTU power consumption feature set; The calculation expression of the weighted fusion is: ; Among them, C is a composite feature vector that integrates time domain and frequency domain information to form a DTU power consumption feature set; T and F represent the feature vectors of the time domain feature set and the frequency domain feature set, respectively; and are the weights of time domain and frequency domain features respectively.

[0024] More specifically, for example, the time domain feature sequence of the DTU in the above example represents the average power consumption change every day of the week, with a total of 7 data points. Fast Fourier transform is applied to transform this sequence to explore its potential periodic behavior. Through calculation, the spectrum of the 7-point sequence is obtained, and it is found that there is a significant peak at the third frequency point, which means that there is a power consumption fluctuation pattern that repeats approximately every 3 days, which may be related to some periodic production activities or environmental factors. Based on this, the power spectral density of this frequency point is selected as one of the frequency domain features, which not only reveals the periodic characteristics of the signal, but also provides a new perspective for subsequent anomaly detection. By combining time domain and frequency domain features, the stacked deep autoencoder model can more comprehensively understand the power consumption behavior of DTU, improve the accuracy of anomaly detection and the ability to recognize complex patterns.

[0025] 103. Based on the historical power consumption characteristics of the DTU, unsupervised encoding-decoding learning is performed on the DTU power consumption characteristic set, and based on a preset first threshold, normal power consumption behavior and abnormal power consumption behavior are identified through reconstruction error analysis; Specifically, the unsupervised encoding-decoding learning is performed on the DTU power consumption feature set based on the historical power consumption features of the DTU, and the normal power consumption behavior and the abnormal power consumption behavior are identified through reconstruction error analysis, including: Based on the historical power consumption characteristics of the DTU, a stacked deep autoencoder model is constructed; Set up 2 hidden layers, each layer includes linear transformation and nonlinear activation function ReLU; Compress and encode the historical power consumption characteristics of the DTU and reconstruct them by decoding; The calculation expression of the hidden layer in the stacked deep autoencoder model is: ; Among them, the The activation value of the layer; It is the nonlinear activation function ReLU; and are the weight matrix and bias vector of the layer respectively; Based on the mean square error, optimize the weights and biases to minimize the reconstruction error; The expression of the mean square error is: ; Where L is the mean square error, which is used to measure the difference between the model output and the actual value; B is the number of samples; is the bth sample input; is the reconstructed output of the bth sample; Input the DTU power consumption feature set into the stacked deep autoencoder model, perform decoding and reconstruction, and output a first reconstruction error; When the first reconstruction error does not exceed a preset first threshold, it is identified as normal power consumption behavior; When the first reconstruction error exceeds a preset first threshold, it is identified as abnormal power consumption behavior.

[0026] More specifically, for example, there is a DTU power consumption feature vector containing 1000 groups We use the training set of 1000 DTUs to build a stacked deep autoencoder model with two hidden layers. The first layer is reduced to half of the original dimension, and the bottleneck layer is kept at 10 dimensions. During the training process, we optimize the weights and biases through the mean square error to minimize the reconstruction error. After training, we input a new set of DTU power consumption feature vectors into the stacked deep autoencoder model. If the reconstruction error of the feature is higher than 99.5% of the samples in the training set, we consider it an abnormal power consumption behavior.

[0027] 104. Calculate the abnormal probability distribution based on the historical power consumption characteristics of the DTU, and adaptively adjust the limit of the first threshold according to the second power consumption information of the monitored DTU; Specifically, the calculating of the abnormal probability distribution based on the historical power consumption characteristics of the DTU, and adaptively adjusting the limit of the first threshold according to the second power consumption information of the monitored DTU include: Based on Gaussian distribution, characterize the historical power consumption characteristics of the DTU and calculate the abnormal probability distribution of the reconstruction error; The calculation formula of the abnormal probability distribution is: ; in, is the reconstruction error; is the abnormal probability distribution, which represents the probability density function of the error and reflects the possibility of the error occurring; and are the mean and standard deviation of historical errors, indicating the central position of the error distribution and the degree of data dispersion; Calculating the dynamic threshold according to the second power consumption information of the DTU; The calculation formula of the dynamic threshold is: ; in, is a factor that controls the strictness of the threshold, and a larger Indicates a higher standard of anomaly detection; ; and are the mean and standard deviation of historical errors, respectively, indicating the central position of the error distribution and the degree of data dispersion.

[0028] More specifically, for example, the stacked deep autoencoder model was used to train and test the normal DTU data of the past year, and 10,000 sets of decoding and reconstruction errors were collected, and the calculated =0.05, =0.1; if you want to control the false alarm rate to 0.5%, you can choose k=3 as a multiple of the standard deviation to calculate the dynamic threshold =0.05+3×0.1=0.35; this means that for a new feature vector (second power consumption information), if its decoding and reconstruction error is greater than 0.35, it is judged as abnormal. This method automatically adjusts the threshold based on historical data, adapts to different time periods or environmental changes, reduces false positives or false negatives caused by improper fixed threshold settings, and improves the flexibility, accuracy, and adaptability of DTU anomaly detection.

[0029] 105. When abnormal power consumption behavior is detected, the preset intelligent early warning system will automatically generate response instructions and adjustment suggestions based on preset rules and strategies.

[0030] Specifically, when abnormal power consumption behavior is detected, the preset intelligent early warning system automatically generates countermeasure instructions and adjustment suggestions according to preset rules and strategies, including: When abnormal power consumption behavior is detected, the intelligent early warning system immediately sends an early warning to the maintenance personnel through at least one method including SMS, email and APP push, and marks the abnormal time period and possible causes; Evaluate the stability of the communication link through indicators such as bit error rate and packet loss rate; Dynamically adjust the operating frequency and power output of the DTU based on the stability of the communication link and task requirements monitored in real time: quickly transmit data when the stability of the communication link is high, and reduce the transmission rate when the stability of the communication link is poor; Adjust the allocation of DTU data transmission tasks to avoid high-load operations during concentrated time periods and reduce abnormal peak power consumption.

[0031] More specifically, for example, a power company has deployed a smart grid monitoring system that collects real-time data from DTUs distributed in various substations and transmits it back to the data center via wireless communication links. The system has been built according to the above method and is now continuously monitoring the power consumption behavior of the DTUs. One day, the smart monitoring system detected that the power consumption of the device numbered DTU-007 increased abnormally between 2 a.m. and 4 a.m., exceeding the normal range by 20%. The system automatically triggers the following responses: Early warning notification: The smart early warning system immediately sends an alert to Xiao Li, the maintenance engineer in charge of the area, via SMS, email, and a dedicated mobile app. The content of the alert includes: "[Smart grid monitoring system] Warning! The power consumption of DTU-007 increased abnormally between 02:00 and 04:00. Possible reasons include external interference or internal hardware failure. Please check in time." Communication link evaluation: The system also checked the communication link status of DTU-007 and found that the bit error rate was 1.5%, which was 0.1% higher than the normal value, and there was slight packet loss. This indicates that the stability of the communication link has decreased, which may affect the efficiency of data transmission.

[0032] Working frequency adjustment: In view of the unstable communication link, the system automatically adjusts the working frequency of DTU-007 from the standard 48MHz to 32MHz to reduce the data transmission rate, reduce the link burden, and reduce power consumption.

[0033] Power output adjustment: Based on the current task priority, the system decides to reduce the frequency of non-urgent data collection, only maintaining real-time monitoring of key parameters such as voltage and current, and further reducing the power output of the DTU to 80% of the normal level.

[0034] Task reallocation: To avoid similar peak power consumption caused by concentrated high-load operations in the future, the system adjusts the data transmission plan and disperses non-urgent data transmission tasks to low-power periods of the day, such as late at night and early morning, to ensure data transmission when power demand is low.

[0035] Through this series of automated adjustments, the power consumption of DTU-007 quickly returned to normal, the communication bit error rate dropped to 0.2%, and the packet loss phenomenon disappeared. Based on the early warning information, the maintenance engineer inspected DTU-007 on site the next morning and found that the external line was slightly worn. He repaired it in time and completely solved the problem.

[0036] In the embodiment of the present invention, the beneficial effect is that the DTU security monitoring method based on power consumption information provided by the embodiment of the present invention, on the one hand, introduces an unsupervised encoding-decoding learning framework, and uses the historical power consumption characteristics of the DTU to train the extracted power consumption feature set. This process can not only learn the complex pattern of normal power consumption behavior, but also automatically distinguish normal and abnormal power consumption behaviors through reconstruction error analysis without manual labeling, and issue early warning responses and strategies in time, which greatly reduces the burden of manual intervention and improves the automation level of monitoring. Combined with the preset first threshold, the framework can flexibly adapt to the normal power consumption range in different scenarios, reducing the probability of false alarms and missed alarms. On the other hand, through in-depth analysis of historical data, calculating the abnormal probability distribution, and dynamically adjusting the limit of the first threshold according to the second power consumption information monitored in real time, this intelligent adjustment mechanism ensures the adaptability and accuracy of the monitoring system, even when dealing with changing complex working conditions and self-generated or environmental conditions, it can maintain a high degree of monitoring efficiency and effectively deal with various unexpected changes.

[0037] The above describes the DTU security monitoring method based on power consumption information in the embodiment of the present invention. The following describes the DTU security monitoring system based on power consumption information in the embodiment of the present invention. Figure 2 ,An embodiment of the DTU safety monitoring system based on power consumption information in the embodiment of the present invention includes: an acquisition module, a processing module and an early warning module; communication connection between each module; Specifically, the acquisition module is used to acquire the power consumption information of the DTU in real time; Specifically, the processing module is used to pre-process the power consumption information of the DTU to obtain a DTU power consumption sequence; extract time domain features and frequency domain features of the DTU power consumption sequence respectively, and fuse them to generate a DTU power consumption feature set; based on the historical power consumption characteristics of the DTU, perform unsupervised encoding-decoding learning on the DTU power consumption feature set, and based on a preset first threshold, identify normal power consumption behavior and abnormal power consumption behavior through reconstruction error analysis; based on the historical power consumption characteristics of the DTU, calculate the abnormal probability distribution, and adaptively adjust the limit of the first threshold according to the second power consumption information of the monitored DTU; and also for executing the above-mentioned DTU security monitoring method based on power consumption information; Specifically, the early warning module is used to automatically generate countermeasure instructions and adjustment suggestions based on preset rules and strategies through a preset intelligent early warning system when abnormal power consumption behavior is detected.

[0038] In the embodiment of the present invention, the beneficial effect is that the DTU security monitoring system based on power consumption information provided by the embodiment of the present invention realizes adaptive, efficient, accurate and timely security monitoring of DTU power consumption behavior, effectively reduces risks, and significantly enhances the security protection capability of the system.

[0039] The present invention also provides a DTU security monitoring device based on power consumption information, the DTU security monitoring device based on power consumption information includes a memory and a processor, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the DTU security monitoring method based on power consumption information in the above-mentioned embodiments.

[0040] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the DTU safety monitoring method based on power consumption information.

[0041] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0042] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0043] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A DTU security monitoring method based on power consumption information, characterized in that: The DTU security monitoring method based on power consumption information includes: Acquire power consumption information of the DTU in real time, and pre-process the power consumption information of the DTU to obtain a DTU power consumption sequence; Extracting time domain features and frequency domain features of the power consumption sequence of the DTU respectively, and fusing them to generate a DTU power consumption feature set; Based on the historical power consumption characteristics of the DTU, unsupervised encoding-decoding learning is performed on the DTU power consumption characteristic set, and based on a preset first threshold, normal power consumption behavior and abnormal power consumption behavior are identified through reconstruction error analysis; Based on the historical power consumption characteristics of the DTU, calculate the abnormal probability distribution, and adaptively adjust the limit of the first threshold according to the second power consumption information of the monitored DTU; When abnormal power consumption behavior is detected, the preset intelligent early warning system will automatically generate response instructions and adjustment suggestions based on preset rules and strategies.

2. The DTU security monitoring method based on power consumption information according to claim 1 is characterized in that: The real-time acquisition of the power consumption information of the DTU and preprocessing the power consumption information of the DTU to obtain a DTU power consumption sequence includes: Through current sensors and voltage sensors, current and voltage are monitored in real time, and power consumption is calculated; Converting the analog signal into a digital signal through an analog-to-digital converter to generate power consumption information of the DTU; Regularly read the power consumption information of the DTU in different states; Performing multi-level filtering on the power consumption information of the DTU to obtain smoothed power consumption information of the DTU; According to the maximum-minimum normalization principle, the power consumption information of the DTU is standardized to obtain a DTU power consumption sequence.

3. The DTU security monitoring method based on power consumption information according to claim 2 is characterized in that: The extracting time domain features and frequency domain features of the power consumption sequence of the DTU respectively and fusing them to generate a DTU power consumption feature set includes: Perform time domain feature extraction on the power consumption sequence of the DTU: The power consumption sequence is converted into a three-dimensional tensor according to [sample quantity, time step, power consumption value feature dimension] and input into the input layer; Set 32 ​​one-dimensional convolution kernels with a size of 3 time steps to slide on the time axis, and convolve to obtain local time series features; The convolution formula is: ; in, is the activation function; Weight is the convolution kernel weight; is a fragment of the input sequence; Bias is the bias term; Set ReLU as the activation function, reduce the time dimension through the maximum pooling layer, and extract key features; Setting a fully connected layer and an output layer to identify or predict a specific power consumption mode or state, outputting a time domain feature sequence of the DTU, and generating a time domain feature set of the DTU; Perform frequency domain feature extraction on the power consumption sequence of the DTU: Performing a fast Fourier transform operation on the time domain characteristic sequence of the DTU to generate a spectrum diagram; The Fourier transform formula is: ; Wherein, X[k] represents the complex value of the transformed frequency domain signal at the kth frequency point; x[n] is the value of the time domain characteristic sequence of the DTU at the nth time point; N is the length of the time domain characteristic sequence of the DTU; It is a complex exponential function and the basic rotation factor in Fourier transform, which is used to decompose the time domain characteristic sequence of DTU into the superposition of different frequency components; Analyze the spectrum graph to identify frequency components and potential frequency patterns based on the frequency and amplitude of periodic power consumption fluctuations; The power spectrum density and frequency peak position of a specific frequency are selected to construct the frequency domain feature set of the DTU.

4. The DTU security monitoring method based on power consumption information according to claim 3 is characterized in that: The extracting time domain features and frequency domain features of the power consumption sequence of the DTU respectively, and fusing to generate a DTU power consumption feature set, further includes: The time domain feature set of the DTU and the frequency domain feature set of the DTU are integrated to generate a DTU power consumption feature set, including: Time-aligning the time domain feature set of the DTU and the frequency domain feature set of the DTU; Performing a weighted fusion operation on the time domain feature set of the DTU and the frequency domain feature set of the DTU to generate the DTU power consumption feature set; The calculation expression of the weighted fusion is: ; Among them, C is a composite feature vector that integrates time domain and frequency domain information to form a DTU power consumption feature set; T and F represent the feature vectors of the time domain feature set and the frequency domain feature set respectively; and are the weights of time domain and frequency domain features respectively.

5. The DTU security monitoring method based on power consumption information according to claim 4 is characterized in that: The performing unsupervised encoding-decoding learning on the DTU power consumption feature set based on the historical power consumption features of the DTU, and identifying normal power consumption behavior and abnormal power consumption behavior through reconstruction error analysis, includes: Based on the historical power consumption characteristics of the DTU, a stacked deep autoencoder model is constructed; Set up 2 hidden layers, each layer includes linear transformation and nonlinear activation function ReLU; Compress and encode the historical power consumption characteristics of the DTU and reconstruct them by decoding; The calculation expression of the hidden layer in the stacked deep autoencoder model is: ; Among them, the The activation value of the layer; It is the nonlinear activation function ReLU; and are the weight matrix and bias vector of the layer respectively; Based on the mean square error, optimize the weights and biases to minimize the reconstruction error; The expression of the mean square error is: ; Where L is the mean square error, which is used to measure the difference between the model output and the actual value; B is the number of samples; is the bth sample input; is the reconstructed output of the bth sample; Input the DTU power consumption feature set into the stacked deep autoencoder model, perform decoding and reconstruction, and output a first reconstruction error; When the first reconstruction error does not exceed a preset first threshold, it is identified as normal power consumption behavior; When the first reconstruction error exceeds a preset first threshold, it is identified as abnormal power consumption behavior.

6. The DTU security monitoring method based on power consumption information according to claim 5 is characterized in that: The calculating the abnormal probability distribution based on the historical power consumption characteristics of the DTU, and adaptively adjusting the limit of the first threshold according to the second power consumption information of the monitored DTU, includes: Based on Gaussian distribution, characterize the historical power consumption characteristics of the DTU and calculate the abnormal probability distribution of the reconstruction error; The calculation formula of the abnormal probability distribution is: ; in, is the reconstruction error; is the abnormal probability distribution, which represents the probability density function of the error and reflects the possibility of the error occurring; and are the mean and standard deviation of historical errors, indicating the central position of the error distribution and the degree of data dispersion; Calculating the dynamic threshold according to the second power consumption information of the DTU; The calculation formula of the dynamic threshold is: ; in, is a factor that controls the strictness of the threshold, and a larger Indicates a higher standard of anomaly detection; ; and are the mean and standard deviation of historical errors, respectively, indicating the central position of the error distribution and the degree of data dispersion.

7. The DTU security monitoring method based on power consumption information according to claim 6 is characterized in that: When abnormal power consumption behavior is detected, the preset intelligent early warning system automatically generates countermeasure instructions and adjustment suggestions according to preset rules and strategies, including: When abnormal power consumption behavior is detected, the intelligent early warning system immediately sends an early warning to the maintenance personnel through at least one method including SMS, email and APP push, and marks the abnormal time period and possible causes; Evaluate the stability of the communication link through indicators such as bit error rate and packet loss rate; Dynamically adjust the operating frequency and power output of the DTU based on the stability of the communication link and task requirements monitored in real time: quickly transmit data when the stability of the communication link is high, and reduce the transmission rate when the stability of the communication link is poor; Adjust the allocation of DTU data transmission tasks to avoid high-load operations during concentrated time periods and reduce abnormal peak power consumption.

8. A DTU safety monitoring system based on power consumption information, characterized in that: The DTU safety monitoring system based on power consumption information includes: An acquisition module, used to acquire the power consumption information of the DTU in real time; A processing module, used to pre-process the power consumption information of the DTU to obtain a DTU power consumption sequence; extract time domain features and frequency domain features of the DTU power consumption sequence respectively, and fuse them to generate a DTU power consumption feature set; based on the historical power consumption characteristics of the DTU, perform unsupervised encoding-decoding learning on the DTU power consumption feature set, and based on a preset first threshold, identify normal power consumption behavior and abnormal power consumption behavior through reconstruction error analysis; based on the historical power consumption characteristics of the DTU, calculate the abnormal probability distribution, and adaptively adjust the limit of the first threshold according to the second power consumption information of the monitored DTU; and also used to execute the DTU security monitoring method based on power consumption information as described in any one of claims 1-7; The early warning module is used to automatically generate response instructions and adjustment suggestions based on preset rules and strategies through the preset intelligent early warning system when abnormal power consumption behavior is detected.

9. A DTU safety monitoring device based on power consumption information, characterized in that: The DTU safety monitoring device based on power consumption information includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instruction in the memory so that the DTU security monitoring device based on power consumption information executes the DTU security monitoring method based on power consumption information as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by the processor, the DTU security monitoring method based on power consumption information is implemented as described in any one of claims 1-7.

Citation Information

Patent Citations

  • FPGA dynamic power consumption estimation method based on BP neural network

    CN104199536A

  • Power grid embedded terminal safety monitoring method and system based on side channel

    CN111030299A

  • Method for constructing anomaly detection model, anomaly detection method and anomaly detection system

    CN117892214A

  • Miniaturized low-power-consumption analog power closed-loop system and method in MIMO-OFDM system

    CN118574205A

  • Performance of Services Based On Power Consumption

    US20160034024A1

Cited By

  • Low-power-consumption optimization method, device and equipment for MEMS sensor and storage medium

    CN119916884A

  • Low-Power Optimization Method, Device, Equipment and Storage Medium for MEMS Sensors

    CN119916884B