A millimeter-wave radar system and method for monitoring icing on power transmission lines

Abstract

This invention relates to the field of power system safety monitoring technology, and proposes a millimeter-wave radar transmission line icing monitoring system and method, capable of all-weather icing monitoring under complex meteorological conditions (such as rain, snow, fog, etc.). By extracting frequency domain features through Fast Fourier Transform (FFT) and combining it with algorithms such as Support Vector Regression (SVR) to invert icing thickness, the system can obtain the icing thickness on transmission lines in real time and accurately. This invention utilizes the wide coverage characteristics of millimeter-wave radar to monitor large areas of transmission lines. Compared to locally installed fiber optic sensors, this system can cover a wider line area with a smaller number of radar devices, extending the monitoring range to tens of kilometers, greatly reducing equipment installation and maintenance costs. When the icing thickness of the transmission line or the rate of increase in icing thickness exceeds a set safety threshold, this invention triggers an alarm to remind maintenance personnel to clean the line.

Description

Technical Field

[0001] This invention relates to the field of power system safety monitoring technology, and in particular to a millimeter-wave radar transmission line icing monitoring system and method. Background Technology

[0002] In power transmission systems, the safe and stable operation of transmission lines is crucial for power supply. However, transmission lines are typically erected outdoors and exposed to harsh weather conditions for extended periods. When encountering low temperatures and high humidity, these lines are highly susceptible to icing. Icing increases the weight of transmission lines, potentially leading to serious accidents such as line breaks and tower collapses. Furthermore, icing increases the resistance of transmission lines, resulting in decreased transmission efficiency. Therefore, monitoring and early warning systems for icing on transmission lines are extremely important for the power industry.

[0003] In existing technologies, the main methods for detecting icing on transmission lines include the following:

[0004] 1. Traditional manual inspection method: The icing condition of transmission lines is judged by inspection or infrared thermal imaging technology. This method has the disadvantages of low efficiency, poor real-time performance and limitation by environmental conditions. Moreover, the icing of transmission lines is caused by severe weather, and manual inspection under severe weather conditions is prone to unexpected situations.

[0005] 2. Video surveillance and image recognition technology: Some systems use video surveillance equipment to monitor power transmission lines in real time and combine image processing algorithms to analyze icing conditions; however, video surveillance systems are prone to failure in low visibility weather conditions such as clouds, fog, rain, and snow, and their coverage is limited, making it difficult to meet the detection needs of large-scale power transmission lines.

[0006] 3. Fiber optic sensing technology: By installing fiber optic sensors on transmission lines, line deformation caused by icing can be detected; however, fiber optic sensors are complex to install and have high maintenance costs, and because they can only monitor local areas, they cannot cover the entire transmission line.

[0007] While these existing technologies have addressed the icing detection problem to some extent, they still have limitations and shortcomings in practical applications. These include poor real-time performance (manual and video surveillance rely on weather and personnel conditions, making 24 / 7 monitoring impossible), limited monitoring range (traditional methods struggle to comprehensively monitor power transmission lines over vast areas, resulting in blind spots), poor adaptability to severe weather (the detection effectiveness of video surveillance and fiber optic sensors significantly decreases in adverse weather conditions such as rain, snow, and fog), and high cost and complex maintenance (the installation and maintenance of fiber optic sensors and other equipment are expensive, increasing the overall operational complexity).

[0008] Existing icing monitoring technologies are insufficient to fully meet the needs of practical applications, especially under severe weather conditions. Traditional methods are ineffective and fail to provide efficient and accurate real-time monitoring. Therefore, there is an urgent need for a transmission line icing monitoring system that can operate stably and effectively under various weather conditions and has high precision and low cost.

[0009] Based on this, the present invention proposes a method and system for monitoring icing of transmission lines based on millimeter-wave radar. By utilizing the penetration capability and high-resolution characteristics of millimeter-wave radar and combining it with signal processing technology, the shortcomings of the prior art are overcome, and real-time monitoring and early warning of icing conditions of transmission lines can be achieved under severe weather conditions. Summary of the Invention

[0010] The purpose of this invention is to address the shortcomings of existing technologies by providing a millimeter-wave radar transmission line icing monitoring system and method. By utilizing the penetrating power and high-resolution characteristics of millimeter-wave radar, combined with signal processing technology, the invention overcomes the deficiencies of existing technologies and enables real-time monitoring and early warning of transmission line icing under severe weather conditions.

[0011] To achieve the above objectives, the present invention adopts the following technical solution:

[0012] A millimeter-wave radar transmission line icing monitoring system includes: a millimeter-wave radar module, a signal processing module, an icing thickness inversion module, a real-time monitoring module, a communication module, and a visualization module;

[0013] The millimeter-wave radar module is used to transmit millimeter-wave electromagnetic signals and receive echo signals reflected from the surface of the transmission line, and output the echo signals reflected from the surface of the transmission line to the signal processing module.

[0014] The signal processing module is used to receive the echo signal output by the millimeter-wave radar module, preprocess the echo signal and extract the backscattering coefficient, phase angle and amplitude of the echo signal, and output the backscattering coefficient, phase angle and amplitude of the echo signal to the ice thickness inversion module.

[0015] The ice thickness inversion module is used to output the ice thickness on the transmission line based on the backscattering coefficient, phase angle and amplitude of the echo signal output by the signal processing module, combined with the ice thickness inversion model, and output the ice thickness data on the transmission line to the real-time monitoring module.

[0016] The real-time monitoring module is used to determine whether there is abnormal icing based on the icing thickness data output by the icing thickness inversion module. If there is abnormal icing, an alarm is triggered, and the alarm information and icing thickness data are output to the communication module.

[0017] The communication module is used to transmit alarm information and ice thickness data to the remote monitoring center in real time via wireless communication, so that power system managers can monitor and analyze the data.

[0018] The visualization module is used to display alarm information and icing thickness data.

[0019] Preferably, the millimeter-wave radar module is mounted on a drone or other flight platform.

[0020] Preferably, the ice thickness inversion model is obtained based on the following method:

[0021] Step 1: Data Acquisition; Collect echo signals from transmission lines under different icing thicknesses to form a sample set;

[0022] Step 2: Data preprocessing; cleaning and standardizing the acquired echo signals;

[0023] Step 3: Echo signal feature extraction; extract the backscattering coefficient, phase angle, and amplitude of the preprocessed echo signal; assemble the backscattering coefficient, frequency domain phase, and amplitude information features into a feature vector, which is used as the input to the subsequent model;

[0024] Step 4: Train the ice thickness inversion model; Using the support vector regression model, train an ice thickness inversion model with feature vectors as input and ice thickness as output. Cross-validation is used to optimize hyperparameters during training.

[0025] Step 5: Use the validation set to test the inversion effect of the icing thickness inversion model; set the target accuracy of the test set to be above 95%.

[0026] Preferably, the feature extraction in step 3 includes the following steps:

[0027] The echo signal is converted into backscattering coefficients through time-frequency analysis and radar equations, as shown in the following formula:

[0028] ;

[0029] In the formula It is the received signal power. Transmitted signal power It is the distance at which the radar reaches the target. , It refers to the gain of the transmitting and receiving antennas. It is the radar wavelength;

[0030] The preprocessed time-domain echo signal is subjected to Fourier transform to generate a frequency-domain echo signal, and the phase angle and amplitude information of the echo signal are extracted from the spectrum.

[0031] Preferably, the objective function of the support vector regression model in step 4 is as follows:

[0032] ;

[0033] in For feature vectors, As weight, It is a nonlinear mapping function. This is the bias; the optimal value is found by minimizing the loss function during training. and .

[0034] Preferably, after step 5 is completed, actual field operation data is collected periodically, and the current inversion model is retrained using this data; the existing model is slightly updated using new data, and the existing model weights or support vectors are fine-tuned using the newly added data.

[0035] Preferably, the real-time monitoring module analyzes the changes in the ice thickness of the transmission line in real time during the monitoring process, and triggers an alarm when the ice thickness of the transmission line or the rate of increase of the ice thickness exceeds a set safety threshold.

[0036] Another objective of this invention is to provide a millimeter-wave radar method for monitoring icing on power transmission lines, using the aforementioned millimeter-wave radar system for monitoring icing on power transmission lines, comprising the following steps:

[0037] S1, Inspection begins; the drone's millimeter-wave radar module begins its inspection along the power transmission line, collecting echo signals in real time;

[0038] S2. Signal Processing: The signal processing module preprocesses the acquired echo signals to remove noise and interference, and extracts features.

[0039] S3. Data Analysis: Transmit the accretion information to the ice thickness inversion module, and calculate the ice thickness through the ice thickness inversion model;

[0040] S4. Monitoring and Alarm: The real-time monitoring module updates the icing thickness data of the transmission line in real time, performs icing trend analysis of the transmission line, calculates the growth rate of icing thickness of the transmission line, and once it exceeds the preset threshold, the real-time monitoring module automatically triggers an alarm and sends information to the monitoring center through the communication module.

[0041] S5. Data transmission and feedback: All monitoring data and alarm information are transmitted to the remote monitoring center in real time, and maintenance personnel can obtain information through a visual interface.

[0042] Preferably, step S2 includes the following steps:

[0043] S2.1 Noise Reduction Processing: Bandpass filters are used to remove noise outside the frequency band, and wavelet transform is used to remove high-frequency noise in the echo signal; the signal amplitude is normalized to standardize the signal in each time period.

[0044] S2.2 Feature Extraction; Extracted features include backscattering coefficients, frequency domain phase, and amplitude information;

[0045] The backscattering coefficient is derived from the echo signal through time-frequency analysis and radar equations. The specific formula is as follows:

[0046] ;

[0047] In the formula It is the received signal power. Transmitted signal power It is the distance at which the radar reaches the target. , It refers to the gain of the transmitting and receiving antennas. It is the radar wavelength;

[0048] Frequency domain phase is calculated by extracting frequency domain information from the preprocessed signal using a Fast Fourier Transform, with the specific formula as follows:

[0049] ;

[0050] in, The complex spectrum obtained after FFT;

[0051] Amplitude information is calculated from the signal amplitude, which serves as the intensity feature of the target. The amplitude can be directly extracted from the FFT result, in the following form:

[0052] ;

[0053] S2.3, Feature Vector Assembly: Combine the backscattering coefficients, frequency domain phase, and amplitude information into a feature vector; the number of extracted features is n, and the feature vector is represented as:

[0054] ;

[0055] Where F is the feature vector, σ is the backscattering coefficient, ϕ(fi) is the phase angle of the i-th frequency domain signal, A(fi) is the amplitude information of the i-th frequency domain signal, and n is the number of extracted features, representing the resolution of the signal in the frequency domain.

[0056] The present invention discloses a millimeter-wave radar transmission line icing monitoring system and method, which has the following beneficial effects.

[0057] This invention enables all-weather icing monitoring under complex meteorological conditions (such as rain, snow, and fog). By extracting frequency domain features through Fast Fourier Transform (FFT) and combining it with algorithms such as Support Vector Regression (SVR) to invert icing thickness, it can obtain the icing thickness on transmission lines in real time and accurately. Utilizing the wide coverage of millimeter-wave radar, this invention can monitor large areas of transmission lines. Compared to locally installed fiber optic sensors, this system can cover a wider area of ​​lines with a smaller number of radar devices, extending the monitoring range to tens of kilometers, greatly reducing equipment installation and maintenance costs. When the icing thickness of the transmission line or the rate of increase in icing thickness exceeds a set safety threshold, this invention triggers an alarm to remind maintenance personnel to clean the line. Attached Figure Description

[0058] Figure 1 This is a schematic diagram of the structure of the millimeter-wave radar transmission line icing monitoring system of the present invention.

[0059] Figure 2 This is a flowchart of the millimeter-wave radar transmission line icing monitoring method of the present invention.

[0060] In the attached diagram: 1. Wearable part; 11. Back buckle; 2. Safety rope; 3. Hook; 4. Monitoring main unit; 5. Monitoring auxiliary unit; 51. Array displacement sensor; 6. Handheld monitoring terminal. Detailed Implementation

[0061] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0062] In this document, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The term "embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment, nor does it specifically limit its independence or connection with other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the technical features mentioned in each embodiment can be combined in any way to form corresponding implementable technical solutions.

[0063] Unless otherwise defined, the technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the use of related terms herein is merely for the purpose of describing particular embodiments and is not intended to limit this application.

[0064] In the description of this application, the term "and / or" is used to describe the logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and A and B exist simultaneously. Additionally, the character " / " in this document generally indicates that the preceding and following objects have an "or" logical relationship.

[0065] Unless otherwise expressly specified or limited, the terms "installation," "connection," "linking," "fixing," and "setting," as used in the description of the embodiments of this application, should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal connection of two components or the interaction between two components. For those skilled in the art to which this application pertains, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0066] Example 1

[0067] Please refer to Figure 2 A millimeter-wave radar transmission line icing monitoring system includes: a millimeter-wave radar module, a signal processing module, an icing thickness inversion module, a real-time monitoring module, a communication module, and a visualization module;

[0068] A millimeter-wave radar module is used to transmit millimeter-wave electromagnetic signals and receive echo signals reflected from the surface of the power transmission line, and output the echo signals reflected from the surface of the power transmission line to a signal processing module. In this application, a SAR device is carried by a UAV or other flight platform to conduct inspections above the power transmission line, providing wide-area coverage of the entire line. The SAR radar signal acquisition module transmits electromagnetic waves and receives reflected signals from the surface of the power transmission line and the surrounding environment. Especially when ice forms or foreign objects adhere, the reflection intensity and scattering characteristics of the signal will change. The echo data acquired by the SAR radar has high spatial resolution and can accurately depict the geometry and surface condition of the power transmission line.

[0069] The signal processing module is used to receive the echo signal output by the millimeter-wave radar module, preprocess the echo signal and extract the backscattering coefficient, phase angle and amplitude of the echo signal, and output the backscattering coefficient, phase angle and amplitude of the echo signal to the ice thickness inversion module.

[0070] The icing thickness inversion module is used to output the icing thickness on the transmission line based on the backscattering coefficient, phase angle and amplitude of the echo signal output by the signal processing module, combined with the icing thickness inversion model, and output the icing thickness data on the transmission line to the real-time monitoring module.

[0071] The real-time monitoring module is used to determine whether there is abnormal icing based on the icing thickness data output by the icing thickness inversion module. If there is abnormal icing, an alarm is triggered, and the alarm information and icing thickness data are output to the communication module.

[0072] In this embodiment, the preferred real-time monitoring module analyzes the changes in the ice thickness of the transmission line in real time during the monitoring process. When the ice thickness or the rate of increase of the ice thickness exceeds a set safety threshold, an alarm is triggered. In this application, when setting the threshold, historical ice thickness data of the transmission line is collected to construct a sample library. The distribution characteristics of the ice thickness in the sample library are analyzed to identify the normal and abnormal ranges. The mean, standard deviation, and extreme values ​​of the sample data are calculated to establish a statistical model of the ice thickness. Based on the statistical results, a normal thickness upper limit threshold (Tupper) is set, which is the mean plus 1 to 2 standard deviations, representing the upper limit of the normal ice thickness. An abnormal thickness alarm threshold (Talarm) is the mean plus 3 standard deviations; if the ice thickness exceeds this value, it is considered abnormal. The growth rate threshold is set by analyzing the rate of change of ice thickness in historical data, calculating the average growth rate over different time intervals (e.g., 1 hour, 3 hours, and 24 hours), and setting a growth rate threshold (Vthreshold). The calculation formula is as follows:

[0073] Vthreshold = mean(Vhistorical) + k⋅std(Vhistorical); where k is a user-defined empirical value, ranging from 1 to 2;

[0074] Once the detected ice thickness or its growth rate exceeds the set threshold, the real-time monitoring module will trigger an alarm according to the following strategy:

[0075] When the ice thickness reaches the upper limit threshold of the normal thickness (Tupper), the system issues an alarm signal indicating a possible ice accumulation anomaly, prompting monitoring personnel to pay attention.

[0076] When the ice thickness exceeds the abnormal thickness alarm threshold Talarm or the growth rate exceeds Vthreshold, the system automatically triggers an ice abnormality and immediately notifies relevant maintenance personnel to conduct on-site inspection.

[0077] The communication module is used to transmit alarm information and ice thickness data to the remote monitoring center in real time via wireless communication, so that power system managers can monitor and analyze the data.

[0078] The visualization module displays alarm information and icing thickness data. In this embodiment, the visualization module presents the data in the form of visual charts, including real-time thickness change curves, historical data comparisons, and trend prediction graphs. Users can intuitively view the icing status of the line through the interface. Alarm prompts pop up on the user interface, displaying abnormal icing areas and providing specific information such as time and thickness to help maintenance personnel handle the situation promptly.

[0079] In this embodiment, the millimeter-wave radar module is preferably mounted on a drone or other flight platform.

[0080] In this embodiment, the preferred ice thickness inversion model is obtained based on the following method:

[0081] Step 1: Data Acquisition; Collect echo signals from transmission lines under different icing thicknesses to form a sample set;

[0082] Step 2: Data preprocessing; The collected echo signals are cleaned and standardized; In this application, the echo signals are converted into digital signals by a high-speed analog-to-digital converter, and wavelet transform and adaptive filters are used to remove environmental noise and non-target signals. Normalization is used to reduce the signal amplitude variation in different time periods.

[0083] Step 3: Echo signal feature extraction; extract the backscattering coefficient, phase angle, and amplitude of the preprocessed echo signal; assemble the backscattering coefficient, frequency domain phase, and amplitude information features into a feature vector, which serves as the input to the subsequent model; define the target variable as the icing thickness H on the transmission line, calibrate it through field measurements, and establish the mapping relationship between the features and the target variable;

[0084] In this embodiment, the feature extraction in step 3 preferably includes the following steps:

[0085] The echo signal is converted into backscattering coefficients through time-frequency analysis and radar equations, as shown in the following formula:

[0086] ;

[0087] In the formula It is the received signal power. Transmitted signal power It is the distance at which the radar reaches the target. , It refers to the gain of the transmitting and receiving antennas. It is the radar wavelength;

[0088] The preprocessed time-domain echo signal is subjected to a Fourier transform to generate a frequency-domain echo signal. The phase angle and amplitude information of the echo signal are extracted from the spectrum. In this embodiment, the main calculation process of the Fourier transform (FFT) involves dividing a long time-domain signal sequence into two subsequences at even and odd positions. The Fourier transforms of these two subsequences are calculated recursively. The Fourier transform results of the even and odd sequences are weighted and combined to obtain the Fourier transform result of the original sequence. The purpose of this processing is to preserve as many signal characteristics as possible related to icing as possible, while filtering out irrelevant information such as noise and interference. The extracted features provide a reliable data foundation for subsequent icing thickness inversion.

[0089] Step 4: Train the ice thickness inversion model; Using the support vector regression model, train a training ice thickness inversion model with feature vectors as input and ice thickness as output. Cross-validation is used to optimize hyperparameters during training; SVR has good generalization ability and is suitable for handling nonlinear problems; Divide the collected sample data into training set, validation set and test set in a ratio of 7:2:1; Use the training set to train the SVR model;

[0090] In this embodiment, the objective function of the support vector regression model in preferred step 4 is as follows:

[0091] ;

[0092] in For feature vectors, As weight, It is a nonlinear mapping function. This is the bias; the optimal value is found by minimizing the loss function during training. and ;

[0093] Step 5: Use the validation set to test the inversion effect of the icing thickness inversion model; set the target accuracy of the test set to be above 95%.

[0094] In this embodiment, after step 5 is completed as a preferred step, actual on-site operational data is collected periodically. This data is then used to retrain the current inversion model using a dynamic model update module. New data is used to make minor updates to the existing model, and the newly added data is used to fine-tune the existing model weights or support vectors. The monitored actual icing conditions (e.g., manually verified data) are compared with the prediction results of the icing thickness inversion model. If the prediction results of the icing thickness inversion model deviate significantly from the actual situation, the weights of the icing thickness inversion model are automatically updated to ensure that the icing thickness inversion model always accurately reflects the actual on-site situation. When a large fluctuation or abnormal trend is detected in the icing monitoring results (e.g., a sudden increase or decrease in icing thickness), an anomaly detection mechanism is triggered, and the icing thickness inversion model update program is automatically started. Anomaly detection identifies situations where the output of the icing thickness inversion model does not match the current environment by monitoring the changing trends of the data in real time. Subsequently, the icing thickness inversion model is corrected through adaptive adjustment or model retraining to adapt to the new environmental conditions.

[0095] Example 2

[0096] Based on Example 1, please refer to Figure 1 This embodiment provides a millimeter-wave radar method for monitoring icing on power transmission lines, using the millimeter-wave radar system for monitoring icing on power transmission lines described in Embodiment 1, and includes the following steps:

[0097] S1. Inspection begins; the UAV millimeter-wave radar module begins to inspect along the power transmission line and collect echo signals in real time; in this application, GIS is used to plan the inspection route of the UAV to ensure coverage of all power transmission lines. During the flight of the UAV, the SAR device emits millimeter-wave signals, collects echo signals reflected from the power transmission line and its surrounding environment, records data, and outputs the echo signals to the signal processing module.

[0098] S2. Signal Processing: The signal processing module preprocesses the acquired echo signals to remove noise and interference, and extracts features.

[0099] In this embodiment, step S2 preferably includes the following steps:

[0100] S2.1 Noise Reduction Processing: Bandpass filters are used to remove noise outside the frequency band, and wavelet transform is used to remove high-frequency noise in the echo signal; the signal amplitude is normalized to standardize the signal in each time period.

[0101] S2.2 Feature Extraction; Extracted features include backscattering coefficients, frequency domain phase, and amplitude information;

[0102] The backscattering coefficient is derived from the echo signal through time-frequency analysis and radar equations. The specific formula is as follows:

[0103] ;

[0104] In the formula It is the received signal power. Transmitted signal power It is the distance at which the radar reaches the target. , It refers to the gain of the transmitting and receiving antennas. It is the radar wavelength;

[0105] Frequency domain phase is calculated by extracting frequency domain information from the preprocessed signal using a Fast Fourier Transform, with the specific formula as follows:

[0106] ;

[0107] in, The complex spectrum obtained after FFT;

[0108] Amplitude information is calculated from the signal amplitude, which serves as the intensity feature of the target. The amplitude can be directly extracted from the FFT result, in the following form:

[0109] ;

[0110] S2.3, Feature Vector Assembly: Combine the backscattering coefficients, frequency domain phase, and amplitude information into a feature vector; the number of extracted features is n, and the feature vector is represented as:

[0111] ;

[0112] Where F is the eigenvector, σ is the backscattering coefficient, ϕ(fi) is the phase angle of the i-th frequency domain signal, A(fi) is the amplitude information of the i-th frequency domain signal, n is the number of extracted features, representing the resolution of the signal in the frequency domain, T represents the transpose operation, indicating that the eigenvector F is a column vector;

[0113] S3. Data Analysis: The feature information is transmitted to the ice thickness inversion module, which calculates the ice thickness using the ice thickness inversion model. A Kalman filter is used to filter the ice thickness at each time point. The Kalman filter recursively calculates the optimal estimate of the current state. Combining historical data and current measurements, a more accurate thickness prediction is provided.

[0114] S4. Monitoring and Alarm: The real-time monitoring module updates the icing thickness data of the transmission line in real time, performs icing trend analysis of the transmission line, calculates the growth rate of icing thickness of the transmission line, and once it exceeds the preset threshold, the real-time monitoring module automatically triggers an alarm and sends information to the monitoring center through the communication module.

[0115] S5. Data transmission and feedback: All monitoring data and alarm information are transmitted to the remote monitoring center in real time, and maintenance personnel can obtain information through a visual interface.

[0116] Example 3

[0117] Based on Embodiments 1-2, in this embodiment, a millimeter-wave radar transmission line icing monitoring system is provided. The icing thickness inversion module is equipped with a foreign object identification module. The foreign object identification module is used to identify foreign objects attached to the transmission line according to the echo data output by the millimeter-wave radar module after the signal processing module. The identified foreign object information is output to the real-time monitoring module. The real-time monitoring module transmits the foreign object information to the communication module, and the communication module transmits the foreign object information to the monitoring center.

[0118] In this embodiment, the signal processing module extracts the geometric feature information of the foreign object and outputs it to the ice thickness inversion module;

[0119] The icing thickness inversion module is used to output the icing thickness and foreign object attachment information on the transmission line based on the backscattering coefficient, phase angle, and amplitude of the echo signal output by the signal processing module, as well as the geometric characteristics of the foreign object, combined with the icing thickness inversion model.

[0120] In this embodiment, based on Embodiment 1, the icing thickness inversion model outputs information on the icing thickness and foreign object attachment on the transmission line through shared features and multi-task learning. Specifically, it includes the following steps:

[0121] Step 1, Feature Selection: In the data preprocessing stage, distinctive features are extracted using methods such as clustering and principal component analysis (PCA) to jointly construct a more representative feature vector for ice thickness and foreign matter attachment information.

[0122] ;

[0123] Where σice represents features related to ice, and σforeign represents features related to foreign matter;

[0124] Step 2, Joint Model: Use a shared neural network front end, and then use two different output branches, one for predicting ice thickness and the other for foreign object detection;

[0125] During training, the loss function is a composite loss function, consisting of two parts: mean squared error and cross-entropy, which are used for ice thickness inversion and foreign object identification, respectively, as detailed below:

[0126] Where Lice is the loss for ice thickness, Lforeign is the loss for foreign object recognition, and λ is a weighting coefficient used to balance the influence of the two tasks.

[0127] Step 3, Dataset and Training; Dataset preparation: During training, the dataset contains labeled ice thickness and foreign object types, and multiple measurements are taken and data is collected within the same time period; The dataset is divided into training set, validation set and test set in a ratio of 7:2:1;

[0128] An alternating training approach is used, optimizing the loss for ice thickness and foreign object recognition in each training round. After training, the ice thickness inversion model output includes: ice thickness and foreign object type (such as tree branches, plastic bags, etc.). The ice thickness inversion model ensures that foreign objects are not confused with ice.

[0129] Example 4

[0130] Based on Embodiments 1-3, this embodiment provides a millimeter-wave radar method for monitoring icing on transmission lines, using the millimeter-wave radar transmission line icing monitoring system from Embodiments 1-3, including the following steps:

[0131] S1. Inspection begins; the UAV millimeter-wave radar module begins to inspect along the power transmission line and collect echo signals in real time; in this application, GIS is used to plan the inspection route of the UAV to ensure coverage of all power transmission lines. During the flight of the UAV, the SAR device emits millimeter-wave signals, collects echo signals reflected from the power transmission line and its surrounding environment, records data, and outputs the echo signals to the signal processing module.

[0132] S2. Signal Processing: The signal processing module preprocesses the acquired echo signals to remove noise and interference, and extracts features.

[0133] S3. Data Analysis: Transmit the information to the ice thickness inversion module, calculate the ice thickness and identify foreign objects using the ice thickness inversion model;

[0134] S4. Monitoring and Alarm: The real-time monitoring module updates the ice thickness data and foreign object attachment information of the transmission line in real time, performs icing trend analysis of the transmission line, calculates the growth rate of ice thickness, and automatically triggers an alarm once the preset threshold is exceeded and sends information to the monitoring center through the communication module; at the same time, if foreign object attachment information is found, the real-time monitoring module sends foreign object category information to the monitoring center through the communication module.

[0135] S5. Data transmission and feedback: All monitoring data and alarm information are transmitted to the remote monitoring center in real time, and maintenance personnel can obtain information through a visual interface.

[0136] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Substitutions may include replacements of some structures, devices, or method steps, or may be complete technical solutions. Equivalent substitutions or modifications made to the technical solutions and inventive concepts of the present invention should all be covered within the scope of protection of the present invention.

Claims

1. A millimeter-wave radar transmission line icing monitoring system, characterized in that, include: Millimeter-wave radar module, signal processing module, ice thickness inversion module, real-time monitoring module, communication module, and visualization module; The millimeter-wave radar module is used to transmit millimeter-wave electromagnetic signals and receive echo signals reflected back from the surface of the transmission line, and output the echo signals reflected back from the surface of the transmission line to the signal processing module. The signal processing module is used to receive the echo signal output by the millimeter-wave radar module, preprocess the echo signal and extract the backscattering coefficient, phase angle and amplitude of the echo signal, and output the backscattering coefficient, phase angle and amplitude of the echo signal to the ice thickness inversion module. The ice thickness inversion module is used to output the ice thickness on the transmission line based on the backscattering coefficient, phase angle and amplitude of the echo signal output by the signal processing module, combined with the ice thickness inversion model, and output the ice thickness data on the transmission line to the real-time monitoring module. The real-time monitoring module is used to determine whether there is abnormal icing based on the icing thickness data output by the icing thickness inversion module. If there is abnormal icing, an alarm is triggered, and the alarm information and icing thickness data are output to the communication module. The communication module is used to transmit alarm information and ice thickness data to the remote monitoring center in real time via wireless communication, so that power system managers can monitor and analyze the data. The visualization module is used to display alarm information and icing thickness data.

2. The millimeter-wave radar transmission line icing monitoring system as described in claim 1, characterized in that, The millimeter-wave radar module is carried by a drone or other flight platform.

3. The millimeter-wave radar transmission line icing monitoring system as described in claim 1, characterized in that, The ice thickness inversion model was obtained based on the following method: Step 1: Data Acquisition; Collect echo signals from transmission lines under different icing thicknesses to form a sample set; Step 2: Data preprocessing; cleaning and standardizing the acquired echo signals; Step 3: Echo signal feature extraction; Extract the backscattering coefficients, phase angles, and amplitudes of the preprocessed echo signal; assemble the backscattering coefficients, frequency domain phase, and amplitude information features into a feature vector, which serves as the input to the subsequent model; Step 4: Train the ice thickness inversion model; Using the support vector regression model, train an ice thickness inversion model with feature vectors as input and ice thickness as output. Cross-validation is used to optimize hyperparameters during training. Step 5: Use the validation set to test the inversion effect of the icing thickness inversion model; set the target accuracy of the test set to be above 95%.

4. The millimeter-wave radar transmission line icing monitoring system as described in claim 3, characterized in that, The feature extraction in step 3 includes the following steps: The echo signal is converted into backscattering coefficients through time-frequency analysis and radar equations, as shown in the following formula: ； In the formula It is the received signal power. Transmitted signal power It is the distance at which the radar reaches the target. , It refers to the gain of the transmitting and receiving antennas. It is the radar wavelength; The preprocessed time-domain echo signal is subjected to Fourier transform to generate a frequency-domain echo signal, and the phase angle and amplitude information of the echo signal are extracted from the spectrum.

5. The millimeter-wave radar transmission line icing monitoring system as described in claim 3, characterized in that, The objective function of the support vector regression model in step 4 is as follows: ； in For feature vectors, As weight, It is a nonlinear mapping function. For bias; The optimal method is found by minimizing the loss function during training. and .

6. The millimeter-wave radar transmission line icing monitoring system as described in claim 3, characterized in that, After step 5 is completed, collect actual field operation data periodically, and use this data to retrain the current inversion model; use new data to make minor updates to the existing model, and use the newly added data to fine-tune the existing model weights or support vectors.

7. The millimeter-wave radar transmission line icing monitoring system as described in claim 1, characterized in that, The real-time monitoring module analyzes the changes in the ice thickness of the transmission line in real time during the monitoring process. When the ice thickness of the transmission line or the rate of increase of the ice thickness exceeds the set safety threshold, an alarm is triggered.

8. A method for monitoring icing on transmission lines using millimeter-wave radar, comprising the millimeter-wave radar icing monitoring system for transmission lines as described in any one of claims 1 to 7, characterized in that, Includes the following steps: S1, Inspection begins; the drone's millimeter-wave radar module begins its inspection along the power transmission line, collecting echo signals in real time; S2. Signal Processing: The signal processing module preprocesses the acquired echo signals to remove noise and interference, and extracts features. S3. Data Analysis: Transmit the accretion information to the ice thickness inversion module, and calculate the ice thickness through the ice thickness inversion model; S4. Monitoring and Alarm; The real-time monitoring module updates the icing thickness data of the transmission lines in real time, performs icing trend analysis, calculates the growth rate of icing thickness, and automatically triggers an alarm once the preset threshold is exceeded, and sends information to the monitoring center through the communication module. S5. Data transmission and feedback: All monitoring data and alarm information are transmitted to the remote monitoring center in real time, and maintenance personnel can obtain information through a visual interface.

9. The millimeter-wave radar method for monitoring icing on power transmission lines as described in claim 8, characterized in that, Step S2 includes the following steps: S2.1 Noise Reduction Processing: Bandpass filters are used to remove noise outside the frequency band, and wavelet transform is used to remove high-frequency noise in the echo signal; the signal amplitude is normalized to standardize the signal in each time period. S2.2 Feature Extraction; Extracted features include backscattering coefficients, frequency domain phase, and amplitude information; The backscattering coefficient is derived from the echo signal through time-frequency analysis and radar equations. The specific formula is as follows: ； In the formula It is the received signal power. Transmitted signal power It is the distance at which the radar reaches the target. , It refers to the gain of the transmitting and receiving antennas. It is the radar wavelength; Frequency domain phase is calculated by extracting frequency domain information from the preprocessed signal using a Fast Fourier Transform, with the specific formula as follows: ； in, The complex spectrum obtained after FFT; Amplitude information is calculated from the signal amplitude, which serves as the intensity feature of the target. The amplitude can be directly extracted from the FFT result, in the following form: ； S2.3, Feature Vector Assembly: Combine the backscattering coefficients, frequency domain phase, and amplitude information into a feature vector; the number of extracted features is n, and the feature vector is represented as: ； Where F is the feature vector, σ is the backscattering coefficient, ϕ(fi) is the phase angle of the i-th frequency domain signal, A(fi) is the amplitude information of the i-th frequency domain signal, and n is the number of extracted features, representing the resolution of the signal in the frequency domain.

Images