A method and device for monitoring power transmission of a fan unit cable based on thermal imaging

Abstract

The application provides a kind of method and device for monitoring fan unit cable power transmission based on thermal imaging, by utilizing convolutional neural network learning, using thermal imager to collect temperature image of optical cable, pre-processing the collected temperature image information, temperature determination is carried out on the pre-processed temperature image, the transmission power is predicted, through the first early warning, secondary early warning, the function of early warning is realized, the possibility of fire occurrence is avoided, at the same time, the present application can be external on cable, simple and easy to operate, the difficulty of replacement is extremely low.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, and specifically to a method and apparatus for monitoring power transmission of wind turbine cables based on thermal imaging. Background Technology

[0002] Wind turbines convert the kinetic energy of wind into mechanical kinetic energy, and then into electrical kinetic energy—this is wind power generation. The principle of wind power generation is to use wind to drive the rotation of windmill blades, and then use a speed increaser to increase the rotation speed, thereby driving the generator to produce electricity. Wind power generation is becoming increasingly popular worldwide because it does not require fuel and does not produce radiation or air pollution. The equipment required for wind power generation is called a wind turbine generator set. Power cables are generally laid through direct burial, conduit, or ductwork. The State Grid's 500kV cables are all laid in tunnels, 66kV-220kV cables are mainly laid in tunnels and ducts, and 10kV-35kV cables are mainly laid in ducts and direct burial. This makes it difficult to visually inspect the operation of power cables and to monitor and locate faults that occur during operation. Before and during a fault, power cables generally experience a rise in local temperature. Temperature is an important parameter reflecting the operating condition of power cables; therefore, temperature monitoring of power cables is a crucial means of monitoring their operation and ensuring their safe operation.

[0003] In existing technologies, cable trench temperature and smoke alarm systems mainly involve deploying temperature and smoke alarm devices in key monitoring areas within the cable trench. These systems can only mitigate the impact of a fire after the power cable catches fire, and cannot fundamentally prevent or reduce the occurrence of power cable fires. In terms of power transmission detection, photoelectric modules are typically used, but these modules need to be built into the cable, making installation and removal very difficult. Summary of the Invention

[0004] The purpose of this invention is to provide a method and device for monitoring the power transmission of wind turbine cables based on thermal imaging, so as to solve the problem that it can only reduce the scope of the accident after the power cable catches fire, but cannot fundamentally prevent or reduce the occurrence of power cable fire accidents; and that the use of photoelectric modules, which need to be built into the cable, is very difficult to install and remove.

[0005] On one hand, the present invention provides a method for monitoring power transmission in wind turbine cables based on thermal imaging, comprising:

[0006] The optical signal is converted into an electrical signal through the optical-to-electrical processing unit. The data and data results are used to learn the convolutional neural network. Multiple prior information such as the intensity of the light signal in the receiving direction, the intensity of the light signal in the transmitting direction, the self temperature, and the ambient temperature are added to the data model to learn the relationship between temperature and optical signal transmission power. A deep convolutional neural network (DCNN) is used for adaptive learning. All data in the data model should be within the normal range.

[0007] Temperature images of optical cables are acquired using a thermal imager.

[0008] Preprocess the acquired temperature image information;

[0009] Temperature determination is performed on the preprocessed temperature image;

[0010] Predict the transmission power.

[0011] Furthermore, learning is conducted using convolutional neural networks, including:

[0012] The optical signal is converted into an electrical signal through the optical-to-electrical processing unit. The data and data results are used to learn the convolutional neural network. Multiple prior information such as the intensity of the light signal in the receiving direction, the intensity of the light signal in the transmitting direction, the self temperature, and the ambient temperature are added to the data model to learn the relationship between temperature and optical signal transmission power. A deep convolutional neural network (DCNN) is used for adaptive learning. All data in the data model should be within the normal range.

[0013] Furthermore, temperature images of the optical cable are acquired using a thermal imager, including:

[0014] The thermal imager is detachably placed at the location where measurement is required. The thermal imager has night vision capabilities and collects real-time images of the temperature information of the wind turbine cable. The detected area is set as a region, which is composed of multiple sub-regions. The thermal imager is controlled to detect multiple sub-regions in the monitoring area one by one, and the acquisition frequency is set to once every five seconds.

[0015] Furthermore, the acquired temperature image information is preprocessed, including:

[0016] The area temperature image of the wind turbine cable is converted to grayscale. In the RGB model, if R=G=B, then color represents a grayscale color. The value of R=G=B is called the grayscale value. Therefore, each pixel of the grayscale image only needs one byte to store the grayscale value (also known as the intensity value or brightness value).

[0017] The grayscale image of the wind turbine cable area temperature is converted to grayscale. The grayscale conversion includes: grayscale stretching of the wind turbine cable area temperature image so that the grayscale level occupies the entire area of ​​pixel value 0-255.

[0018] Image segmentation is performed on the grayscale processed regional temperature image of the wind turbine cable, dividing the grayscale processed regional temperature image of the wind turbine cable into several specific regions with unique properties;

[0019] Feature extraction is performed on the regional temperature image of the wind turbine cable after image segmentation, and feature extraction is performed on a specific region, namely the region at the wind turbine cable.

[0020] The regional temperature images of the wind turbine cable after feature extraction are normalized.

[0021] Furthermore, temperature determination is performed on the preprocessed temperature image, including:

[0022] The normalized regional temperature image of the wind turbine cable is used to generate a data model. The learned convolutional neural network will compare the data model generated by the actual detection to determine whether the temperature is abnormal.

[0023] If the temperature is abnormal, the alarm unit will issue a level one alarm. It will generate a warning message based on the current time, current location information, the intensity of the light signal in the current receiving direction, the intensity of the light signal in the current sending direction, and the current temperature, and push the warning message to the warning unit.

[0024] If the temperature is normal, the result is predicted using a convolutional neural network.

[0025] Furthermore, the transmission power is predicted, including:

[0026] If the temperature is normal, the power prediction results of the convolutional neural network are analyzed. If the predicted power is within the normal range, no warning is issued.

[0027] If the predicted power exceeds the normal range, the alarm unit will issue a level two warning.

[0028] On the other hand, the present invention provides a device for monitoring power transmission of wind turbine cables based on thermal imaging, comprising:

[0029] Optical-to-electrical processing unit: used to convert optical signals into electrical signals;

[0030] Monitoring unit: Used to monitor the temperature at the location of the wind turbine cable using a thermal imager;

[0031] Transmission unit: Used to transmit data;

[0032] Data processing unit: used to perform image processing operations on all images in a method for monitoring power transmission of wind turbine cables based on thermal imaging;

[0033] Storage unit: Used for data storage;

[0034] Alarm unit: includes: Level 1 warning, Level 2 warning, reset button, and display module;

[0035] The first-level warning is used to warn of abnormal temperature conditions; the second-level warning is used to warn of whether the transmission power is within the normal range; the reset button is used to reset and clear the alarm; the display module is used to display the current time, current location information, current receiving direction light signal strength, and current sending direction light signal strength.

[0036] The beneficial effects of this invention are as follows: This invention provides a method and device for monitoring the power transmission of wind turbine cables based on thermal imaging. By utilizing convolutional neural network learning, a thermal imager is used to acquire temperature images of optical cables. The acquired temperature image information is preprocessed, the temperature of the preprocessed temperature image is determined, and the transmission power is predicted. Through first-level and second-level early warning, the function of early warning is realized, avoiding the possibility of fire. At the same time, this invention can be externally mounted on the cable, is simple to operate, and has extremely low replacement difficulty. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 A flowchart of a method for monitoring power transmission in a wind turbine cable based on thermal imaging, provided by the present invention;

[0039] Figure 2 A schematic diagram of a method for monitoring power transmission in wind turbine cables based on thermal imaging, provided by the present invention;

[0040] Figure 3 This is a schematic diagram of a device for monitoring the power transmission of a wind turbine cable based on thermal imaging, provided by the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The technical solutions provided by various embodiments of this invention will be described in detail below with reference to the accompanying drawings.

[0042] Please see Figures 1 to 2 This invention provides a method for monitoring power transmission in wind turbine cables based on thermal imaging, comprising:

[0043] S101: Learning using a convolutional neural network;

[0044] In this embodiment, convolutional neural network learning is employed, including: converting optical signals into electrical signals using a photoelectric conversion unit; learning the convolutional neural network using sufficient data and data results; incorporating multiple prior information such as the intensity of the received light signal, the intensity of the transmitted light signal, the model's own temperature, and the ambient temperature into the data model to learn the relationship between temperature and optical signal transmission power, thus addressing the complex and variable nature of real-world scenarios and further enhancing the model's adaptability and feature representation capabilities; and using a deep convolutional neural network (DCNN) for adaptive learning. All data in the data model should be within the normal range.

[0045] Learning is achieved through convolutional neural networks (CNNs). A CNN is a feedforward neural network whose artificial neurons can respond to surrounding units within a certain coverage area, demonstrating excellent performance in large-scale image processing. A CNN consists of convolutional layers and pooling layers. CNNs are a highly efficient recognition method that has developed in recent years and attracted widespread attention. This method incorporates multiple prior information into the data model, including the intensity of the received light signal, the intensity of the transmitted light signal, the model's own temperature, and the ambient temperature, to learn the relationship between temperature and light signal transmission power. This addresses the complex and variable nature of real-world scenes, further enhancing the model's adaptability and feature representation capabilities. A deep convolutional neural network (DCNN) is used for adaptive learning, with the extreme temperature set as a threshold.

[0046] To enhance the adaptability and feature representation capabilities of the convolutional neural network (CNN) during training, multiple prior information, such as the intensity of the received and transmitted light signals, its own temperature, and the ambient temperature, are incorporated into the temperature data model. Adaptive learning is achieved using a deep convolutional neural network (DCNN), which addresses the complex and variable nature of real-world scenarios by incorporating these prior information into the temperature data model. In addition to considering static features like color, shape, and contrast, the CNN also fully considers dynamic features such as the intensity of the received and transmitted light signals, its own temperature, and the ambient temperature, improving the accuracy of temperature identification for wind turbine cables. During the initial training phase, the CNN collects as many temperature images or short video samples of wind turbine cables in different scenarios as possible for training. Sample data from specific locations is then refined to optimize the CNN, improving its temperature identification capability and accuracy for wind turbine cables. A threshold temperature is set as the limiting temperature.

[0047] S102: Use a thermal imager to acquire temperature images of the optical cable;

[0048] In this embodiment, the method of acquiring temperature images of the optical cable using a thermal imager also includes:

[0049] The thermal imager can be detachably placed at the location where measurements are needed. It features night vision capabilities. Everything in nature—from Arctic glaciers and flames to the human body and even the frigid depths of space—radiates infrared radiation as long as its temperature is above absolute zero (-273°C). This is due to the thermal motion of molecules within the object. The radiated energy is directly proportional to the fourth power of its temperature, and the wavelength is inversely proportional to its temperature. Infrared imaging technology works by detecting the level of radiated energy of an object. Thermal imaging offers advantages such as rapid response, no disturbance to target temperature, the ability to measure temperature from a distance or for targets that are difficult to access, and a clear view of temperature distribution. Its application at wind turbine cables is a perfect fit, preventing temperature fluctuations from affecting the results; therefore, a thermal imager is chosen.

[0050] The thermal imager acquires real-time images of the temperature information of the wind turbine cable. The detected area is set as a region, which is composed of multiple sub-regions. The thermal imager is controlled to detect each of the multiple sub-regions in the monitoring area, and the acquisition frequency is set to once every five seconds.

[0051] S103: Preprocess the acquired temperature image information;

[0052] In this embodiment, the preprocessing of the acquired temperature image information further includes:

[0053] The regional temperature image of the wind turbine cable was converted to grayscale.

[0054] The grayscale image of the wind turbine cable area temperature is converted to grayscale. The grayscale conversion includes: grayscale stretching of the wind turbine cable area temperature image so that the grayscale level occupies the entire area of ​​pixel value 0-255.

[0055] Image segmentation is performed on the grayscale image of the wind turbine cable's regional temperature. The image is segmented into several specific regions with unique properties. In each region, a seed point is selected as the starting point for growth. Pixels in the neighboring region of the seed point whose similarity to the seed point satisfies a specified growth criterion are searched. These pixels are then merged with the region containing the seed point. The merged pixel is used as the new seed point, and the search and merging process continues until no more new images can be merged. The basic idea of ​​region growing is to group pixels with similar properties to form regions. First, a seed point is selected as the starting point for growth in each region. Then, pixels in the neighboring region of the seed point whose similarity to the seed point satisfies a specified growth criterion are searched and merged with the region containing the seed point. The newly merged pixel is then used as the new seed point, and the search and merging process continues until no more pixels can be merged.

[0056] This paper employs a region growing algorithm to determine candidate character regions. Feature extraction is performed on the regional temperature image of the wind turbine cable after image segmentation. Feature extraction is applied to specific regions, namely the area along the wind turbine cable. Feature extraction transforms group measurements of a certain pattern to highlight representative features of that pattern. This method extracts desired features through image analysis and transformation. Points in the image are divided into different subsets, which often belong to isolated points, continuous curves, or continuous regions. Each pixel is examined to determine whether it represents a feature.

[0057] The regional temperature images of wind turbine cables after feature extraction are normalized because the character sizes in the images vary considerably. Standardizing the size helps improve the accuracy of character recognition and increases the recognition rate, thus facilitating matching with the template. Normalization mainly includes position normalization, size normalization, and stroke thickness normalization.

[0058] S104: Determine the temperature of the preprocessed temperature image;

[0059] In this embodiment, temperature determination of the preprocessed temperature image further includes:

[0060] A data model is generated from the normalized regional temperature images of wind turbine cables. The model is generated using partial least squares (PLS), a mathematical optimization technique that finds the best function match for a set of data by minimizing the sum of squared errors. It uses the simplest method to find some absolutely unknown true values ​​while minimizing the sum of squared errors. Compared to traditional multiple linear regression models, PLS regression has the following advantages: it can perform regression modeling under conditions of severe multicorrelation among independent variables; it allows regression modeling when the number of sample points is less than the number of variables; the final model includes all original independent variables; PLS regression models are easier to identify system information and noise, even some non-random noise; and the regression coefficients of each independent variable are easier to interpret. In this model, every 20 data points are grouped together, and each newly generated data point replaces the previously generated data.

[0061] The learned convolutional neural network will compare the actual detected data model with the actual data model, first determining the temperature and whether the temperature is abnormal.

[0062] If the temperature is abnormal, the alarm unit will issue a level one alarm. It will generate a warning message based on the current time, current location information, the intensity of the light signal in the current receiving direction, the intensity of the light signal in the current sending direction, and the current temperature, and push the warning message to the warning unit.

[0063] If the temperature is normal, the result is predicted using a convolutional neural network.

[0064] S105: Predict the transmission power.

[0065] In this embodiment, predicting the transmission power further includes:

[0066] If the temperature is normal, the power prediction results of the convolutional neural network are analyzed. If the predicted power is within the normal range, no warning is issued.

[0067] If the predicted power exceeds the normal range, the alarm unit will issue a level two warning.

[0068] Please see Figure 3 This invention provides a device for monitoring power transmission in wind turbine cables based on thermal imaging, comprising:

[0069] Optical-to-electrical processing unit 201: used to convert optical signals into electrical signals;

[0070] Monitoring unit 202: Monitors the temperature in the wind turbine cable area using a thermal imager;

[0071] Transmission unit 203: Enables data transmission;

[0072] Data processing unit 204: used to perform image processing operations on all images in a method for monitoring power transmission of wind turbine cables based on thermal imaging;

[0073] Storage unit 205: Used for data storage;

[0074] Alarm unit 206 includes: a first-level warning system, a second-level warning system, a reset button, and a display module;

[0075] The first-level warning is used to alert staff to abnormal temperatures; the second-level warning is used to alert staff to whether the transmission power is within the normal range; the reset button is used to reset and clear the alarm; the display module displays the current time, current location information, current receiving direction light signal strength, and current sending direction light signal strength, allowing staff to understand the cause of the alarm immediately and saving working time.

[0076] This invention provides a method and device for monitoring the power transmission of wind turbine cables based on thermal imaging. The principle is as follows: First, a convolutional neural network is used for learning. A thermal imager is used to acquire temperature images of the optical cable. The acquired temperature images are preprocessed, and then the temperature is determined. If the temperature is abnormal, an alarm unit issues a level one alarm. An early warning message is generated based on the current time, current location, the intensity of the light signal in the current receiving direction, the intensity of the light signal in the current sending direction, and the current temperature, and is pushed to the early warning unit. If the temperature is normal, the convolutional neural network is used to predict the power. If the predicted power is within the normal range, no warning is issued. If the predicted power exceeds the normal range, the alarm unit issues a level two warning. Regardless of whether a level one or level two warning occurs, a reset operation is required. Staff can quickly confirm the warning type, cause, and location based on the information displayed on the module, effectively increasing work efficiency. Through level one and level two warnings, an early warning function is achieved, preventing the possibility of fire.

[0077] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or certain parts of the embodiments of the present invention.

[0078] The above embodiments of the present invention do not constitute a limitation on the scope of protection of the present invention.

Claims

1. A method for monitoring power transmission in wind turbine cables based on thermal imaging, characterized in that, include: The optical signal is converted into an electrical signal through the optical-to-electrical processing unit. The data and data results are used to learn the convolutional neural network. Multiple prior information such as the intensity of the light signal in the receiving direction, the intensity of the light signal in the transmitting direction, the self temperature, and the ambient temperature are added to the data model to learn the relationship between temperature and optical signal transmission power. A deep convolutional neural network (DCNN) is used for adaptive learning. All data in the data model should be within the normal range. Temperature images of optical cables are acquired using a thermal imager. Preprocess the acquired temperature image information; Temperature determination is performed on the preprocessed temperature image; Predict transmission power; The acquired temperature image information is preprocessed, including: The regional temperature image of the wind turbine cable was converted to grayscale. The grayscale image of the wind turbine cable area temperature is converted to grayscale. The grayscale conversion includes: grayscale stretching of the wind turbine cable area temperature image so that the grayscale level occupies the entire area of ​​pixel value 0-255. Image segmentation is performed on the grayscale processed regional temperature image of the wind turbine cable, dividing the grayscale processed regional temperature image of the wind turbine cable into several specific regions with unique properties; Feature extraction is performed on the regional temperature image of the wind turbine cable after image segmentation, and feature extraction is performed on a specific region, namely the region at the wind turbine cable. The regional temperature images of the wind turbine cable after feature extraction are normalized. Temperature determination is performed on the preprocessed temperature image, including: The normalized regional temperature image of the wind turbine cable is used to generate a data model. The learned convolutional neural network will compare the data model generated by the actual detection to determine whether the temperature is abnormal. If the temperature is abnormal, the alarm unit will issue a level one alarm. It will generate a warning message based on the current time, current location information, the intensity of the light signal in the current receiving direction, the intensity of the light signal in the current sending direction, and the current temperature, and push the warning message to the warning unit. If the temperature is normal, the result is predicted based on the convolutional neural network. If the predicted power is within the normal range, no warning is issued. If the predicted power exceeds the normal range, the alarm unit issues a level two warning.

2. A method of monitoring power transfer in a fan bank cable based on thermal imaging according to claim 1, characterized in that, Temperature images of optical cables are acquired using a thermal imager, including: The thermal imager is detachably placed at the location where measurement is required. The thermal imager has night vision capabilities and collects real-time images of the temperature information of the wind turbine cable. The detected area is set as a region, which is composed of multiple sub-regions. The thermal imager is controlled to detect multiple sub-regions in the monitoring area one by one, and the acquisition frequency is set to once every five seconds.

3. A device for monitoring power transmission of a fan unit cable based on thermal imaging, applied to the method for monitoring power transmission of a fan unit cable based on thermal imaging according to any one of claims 1-2, characterized in that, include: Optical-to-electrical processing unit: used to convert optical signals into electrical signals; Monitoring unit: Used to monitor the temperature at the location of the wind turbine cable using a thermal imager; Transmission unit: Used to transmit data; Data processing unit: used for image processing operations according to any one of claims 1 to 2; Storage unit: Used for data storage; Alarm unit: includes: Level 1 warning, Level 2 warning, reset button, and display module; The first-level warning is used to warn of abnormal temperature conditions; the second-level warning is used to warn of whether the transmission power is within the normal range; the reset button is used to reset and clear the alarm; the display module is used to display the current time, current location information, current receiving direction light signal strength, and current sending direction light signal strength.

Images