Safety fault detection system for power generation enterprises based on image recognition

Abstract

A power generation enterprise safety fault detection system based on image recognition relates to the technical field of image recognition; a digital twin model of a power generation device is constructed, a first detection image and a color temperature detection point thereof are obtained, radiation parameters of the color temperature detection point are obtained in combination with the digital twin model, the first detection image is adjusted to obtain a second detection image, heat conduction simulation is performed on the power generation device to obtain a conduction simulation model, a color temperature conduction area of ​​the color temperature detection point is constructed and corresponding color temperature conduction parameters are obtained, the second detection image is adjusted to obtain a third detection image, and whether there is an abnormal color temperature point in the third detection image is judged and feedback is given; the influence of detection angle, surface roughness, and material property value on color temperature recognition can be corrected, which is beneficial to improving the accuracy of safety detection of power generation equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a safety fault detection system for a power generation enterprise based on image recognition. Background Art

[0002] Using image recognition technology to detect safety faults in power generation enterprises is an advanced intelligent solution. It combines image recognition technology with the actual needs of power generation enterprises, aims to improve the safety and reliability of power generation equipment, collect image information of power generation equipment through image sensors, and use advanced image recognition algorithms to process and analyze images, thereby realizing safety fault detection of power generation equipment;

[0003] The most prominent safety inspection of power generation equipment is to detect its temperature and find abnormally high temperature heat source points in time. In the prior art, the detection of heat source points of power generation equipment mostly adopts infrared thermal imaging method, and whether there are abnormally high temperature points in the thermal imaging image is detected to determine whether there are safety problems of the power generation equipment;

[0004] It can be seen that whether the thermal imaging image can accurately reflect the temperature of the power generation equipment will directly affect the results of the safety inspection. However, the existing technology has the following problems: the surface of the power generation equipment often has different influencing factors, and the layout of the thermal imaging equipment is often fixed, which leads to the inevitable error of the thermal imaging image finally formed. However, the existing technology lacks the means to reduce this error;

[0005] Moreover, there are often multiple heat sources in power generation equipment. Due to the superposition of different heat sources, some locations are not overheated themselves but appear overheated in thermal imaging, which also leads to inaccurate safety detection. The prior art also lacks means to solve this problem. In view of the shortcomings of the prior art, the present invention provides a safety fault detection system for power generation enterprises based on image recognition. Summary of the invention

[0006] The purpose of the present invention is to provide a power generation enterprise safety fault detection system based on image recognition.

[0007] The purpose of the present invention can be achieved through the following technical solutions: A power generation enterprise safety fault detection system based on image recognition includes the following modules:

[0008] The equipment twin module is used to collect equipment information of power generation equipment and build a digital twin model of the power generation equipment based on the equipment information;

[0009] An infrared detection module is used to set infrared detection points, obtain an initial detection image of the power generation equipment, obtain a first detection image and several color temperature detection points in the initial detection image, obtain radiation parameters of the color temperature detection points in combination with the digital twin model, adjust the first detection image according to the radiation parameters, and obtain a second detection image;

[0010] A conduction simulation module is used to simulate the heat conduction of the power generation equipment in the digital twin model to obtain a conduction simulation model, construct a color temperature conduction area of ​​each color temperature detection point, and obtain the color temperature conduction parameters of each color temperature detection point, adjust the second detection image according to the color temperature conduction parameters, and obtain a third detection image;

[0011] The safety detection module is used to set a color temperature alarm threshold, determine whether there is an abnormal color temperature point in the third detection image, and provide feedback.

[0012] Furthermore, the process of collecting equipment information of the power generation equipment and constructing a digital twin model of the power generation equipment based on the equipment information includes:

[0013] The equipment information refers to the physical structural parameters of each power generation equipment, including the external shape and internal structure size parameters of the power generation equipment;

[0014] Digital twin technology is used to build digital twin models of each power generation equipment according to the equipment information, and the digital twin models of different power generation equipment are connected according to the connection relationship between the power generation equipment.

[0015] Furthermore, the process of setting infrared detection points, acquiring an initial detection image of the power generation equipment, and acquiring a first detection image and a plurality of color temperature detection points in the initial detection image includes:

[0016] A plurality of infrared detection points are respectively set in the working scene of the power generation equipment, and the surface temperature of the power generation equipment is detected through the infrared detection points and a corresponding initial detection image is generated;

[0017] A color temperature detection threshold is set, and a local area formed by pixels in the initial detection image whose color temperature is greater than or equal to the color temperature detection threshold is used as a first detection image;

[0018] An initial color temperature detection point is obtained in the first detection image, and the probability value of each other pixel point being selected as the next color temperature detection point is obtained. The pixel point corresponding to the maximum probability value is selected as the next color temperature detection point, and this step is repeated until m color temperature detection points are selected.

[0019] Furthermore, the process of obtaining the radiation parameters of the color temperature detection point in combination with the digital twin model, adjusting the first detection image according to the radiation parameters, and obtaining the second detection image includes:

[0020] Upload the initial detection image to the digital twin model of the power generation equipment for synchronization, and obtain the first detection image and the corresponding position of its color temperature detection point in the digital twin model;

[0021] The angle between the line connecting the color temperature detection point and the infrared detection point and the horizontal line is taken as the detection angle S a , obtain the surface roughness S of the color temperature detection point b and material property value S c , the radiation parameters include detection angle, surface roughness, and material property values;

[0022] Get the initial color temperature W of the color temperature detection point c , the initial color temperature of the color temperature detection point is adjusted using the radiation parameter to obtain the first adjusted color temperature W y ;

[0023]

[0024] The radiation parameters of each pixel point in the first detection image are obtained, and the initial color temperature thereof is adjusted to obtain the corresponding first adjusted color temperature, and the second detection image is obtained according to each pixel point under the adjusted first adjusted color temperature.

[0025] Furthermore, the process of simulating the heat conduction of the power generation equipment in the digital twin model to obtain the conduction simulation model includes:

[0026] In the digital twin model, the corresponding density, specific heat capacity and thermal conductivity are set for each component of the power generation equipment, and the corresponding boundary conditions are set for the entire digital twin model. The finite element method is used to simulate the heat conduction of the digital twin model using ANSYS simulation software to obtain a conduction simulation model.

[0027] Furthermore, the process of constructing the color temperature conduction area of ​​each color temperature detection point and obtaining the color temperature conduction parameters of each color temperature detection point includes:

[0028] The first adjusted color temperature of the color temperature detection point is uploaded to the conduction simulation model for synchronization, and the temperature conduction condition of the color temperature detection point at its corresponding position is simulated by the conduction simulation model, and the temperature conduction condition is represented by the color temperature change value at different conduction distances;

[0029] In the conduction simulation model, a color temperature conduction area of ​​the color temperature detection point is constructed with the color temperature detection point as the center, and the color temperature change value on the boundary of the color temperature conduction area is equal to the color temperature of its color temperature monitoring point;

[0030] Constructing a color temperature conduction area of ​​each color temperature detection point based on the second detection image, and obtaining an intersection area between different color temperature conduction areas;

[0031] A color temperature detection point in the intersection area is obtained, and a plurality of color temperature change values ​​of a plurality of color temperature conduction zones in the intersection area at the color temperature detection point are used as color temperature conduction parameters.

[0032] Furthermore, the process of adjusting the second detection image according to the color temperature transmission parameter and obtaining the third detection image includes:

[0033] According to each color temperature conduction area and the intersection area in the second detection image, a color temperature conduction parameter of each pixel point is obtained;

[0034] According to the color temperature transmission parameter of the pixel point, the color temperature W is first adjusted. y Adjust to obtain the second adjusted color temperature W e ;

[0035]

[0036] Among them, W yi is the i-th color temperature transmission parameter B i The first adjusted color temperature of the corresponding color temperature detection point, i=1, 2, ..., n, n represents the number of color temperature conduction areas that produce its intersection area, and the third detection image is obtained according to each pixel point under the second adjusted color temperature after adjustment.

[0037] Furthermore, the process of setting a color temperature alarm threshold and determining whether there is an abnormal color temperature point in the third detection image and providing feedback includes:

[0038] Corresponding color temperature alarm thresholds are set for different power generation equipment, the color temperature of each pixel in the third detection image is compared with its corresponding color temperature alarm threshold, the pixel points that are greater than or equal to the color temperature alarm threshold are judged as abnormal color temperature points, and an alarm signal is generated and fed back.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. The present invention can introduce digital twin technology and image recognition technology into the safety detection of power generation equipment by acquiring a digital twin model and thermal imaging image of the power generation equipment. By acquiring multiple color temperature detection points in the first detection image, the heat source point of the power generation equipment can be effectively obtained;

[0041] 2. By obtaining the radiation parameters of each pixel point and using the radiation coefficient to perform the first adjustment on the color temperature of each pixel point in the first detection image, the influence of the detection angle, surface roughness, and material property value on the color temperature recognition can be corrected, so that the obtained second detection image is more consistent with the actual temperature of the power generation equipment;

[0042] 3. By simulating the heat conduction of the digital twin model of the power generation equipment, the temperature conduction conditions generated by the heat sources at different positions can be obtained, and the color temperature conduction parameters of each pixel point can be obtained. The color temperature conduction parameters can be used to perform a second adjustment on the color temperature distortion at each pixel point, so that the third detection image obtained can effectively reduce the superposition effect of multiple heat sources on a single position, which is conducive to improving the accuracy of safety detection of power generation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0044] like Figure 1 As shown in the figure, the power generation enterprise safety fault detection system based on image recognition includes the following modules:

[0045] The equipment twin module is used to collect equipment information of power generation equipment and build a digital twin model of the power generation equipment based on the equipment information;

[0046] An infrared detection module is used to set infrared detection points, obtain an initial detection image of the power generation equipment, obtain a first detection image and several color temperature detection points in the initial detection image, obtain radiation parameters of the color temperature detection points in combination with the digital twin model, adjust the first detection image according to the radiation parameters, and obtain a second detection image;

[0047] A conduction simulation module is used to simulate the heat conduction of the power generation equipment in the digital twin model to obtain a conduction simulation model, construct a color temperature conduction area of ​​each color temperature detection point, and obtain the color temperature conduction parameters of each color temperature detection point, adjust the second detection image according to the color temperature conduction parameters, and obtain a third detection image;

[0048] The safety detection module is used to set a color temperature alarm threshold, determine whether there is an abnormal color temperature point in the third detection image, and provide feedback.

[0049] It should be further explained that, in the specific implementation process, the process of collecting the equipment information of the power generation equipment and building the digital twin model of the power generation equipment based on the equipment information includes:

[0050] In the embodiment of the present invention, the power generation equipment does not refer to a specific equipment, but includes all different equipment involved in power generation to generate electric energy, and the equipment information refers to the physical structure parameters of each power generation equipment, including the external shape and internal structure size parameters of the power generation equipment;

[0051] Digital twin technology is used to build a digital twin model of each power generation equipment based on the collected equipment information, and the connection relationship between different power generation equipment is obtained. The digital twin models of different power generation equipment are connected according to the obtained connection relationship to obtain the digital twin model of all power generation equipment. At this time, the digital twin model only reflects the physical structure parameters and connection relationship of the power generation equipment.

[0052] It should be further explained that, in a specific implementation process, the process of setting infrared detection points, obtaining an initial detection image of the power generation equipment, and obtaining a first detection image and a plurality of color temperature detection points in the initial detection image includes:

[0053] A plurality of infrared detection points are respectively set in the working scene of the power generation equipment, and an infrared imaging unit is set in each infrared detection point, and the surface temperature of the power generation equipment is detected by the infrared imaging unit, and a corresponding thermal imaging image, i.e., an initial detection image, is generated;

[0054] In the initial detection image, different surface temperatures correspond to colors with different color temperatures, a color temperature detection threshold is set, the color temperature of each pixel in the initial detection image is compared with the set color temperature detection threshold, and a local area composed of pixels greater than or equal to the color temperature detection threshold is marked as a first detection image;

[0055] In the first detection image, the pixel with the highest color temperature is used as the initial color temperature detection point;

[0056] Obtain the shortest distance between each other pixel and the current color temperature detection point, and record the obtained shortest distance as d;

[0057] Obtain the probability value of each other pixel point being selected as the next color temperature detection point, and record the obtained probability value as P;

[0058]

[0059] Wherein, W represents the color temperature of each other pixel;

[0060] The pixel corresponding to the maximum probability value is selected as the next color temperature detection point, and this step is repeated until m color temperature detection points are selected.

[0061] It should be further explained that, in the specific implementation process, the radiation parameters of the color temperature detection point are obtained in combination with the digital twin model, the first detection image is adjusted according to the radiation parameters, and the process of obtaining the second detection image includes:

[0062] The obtained initial detection image is uploaded to the digital twin model of the power generation equipment for synchronization, and the first detection image and the corresponding position of its color temperature detection point are obtained in the digital twin model. Since the infrared detection point is fixed, there are different angles between the positions of different color temperature detection points of the power generation equipment and the infrared detection points, and the surface structures and geometric shapes of different color temperature detection points are also different, which will lead to deviations in the obtained initial detection image.

[0063] Taking any color temperature detection point as an example, the detection angle between the color temperature detection point and its corresponding infrared detection point is obtained in the digital twin model, and the angle between the line connecting the color temperature detection point and the infrared detection point and the horizontal line is taken as the detection angle, and the detection angle is less than or equal to 90 degrees;

[0064] Similarly, the surface roughness and material property value of the color temperature detection point are obtained in the digital twin model. The surface roughness refers to the roughness of the color temperature detection point on the surface of the power generation equipment. The material property value refers to the fixed parameter value corresponding to different preset materials. The radiation parameters include detection angle, surface roughness, and material property value.

[0065] Get the color temperature corresponding to the color temperature detection point, recorded as the initial color temperature W c , and the corresponding detection angle, surface roughness, and material property value are marked as S a , S b , S c , the initial color temperature of the color temperature detection point is adjusted using the radiation parameter to obtain the first adjusted color temperature W y ;

[0066]

[0067] The same method is adopted to respectively obtain the radiation parameters of each pixel point in the first detection image, and the initial color temperature is adjusted to respectively obtain the corresponding first adjusted color temperature, and the second detection image is obtained according to each pixel point under the adjusted first adjusted color temperature. The second detection image has the same area as the first detection image, but a different color temperature.

[0068] It should be further explained that, in the specific implementation process, the process of simulating the heat conduction of the power generation equipment in the digital twin model to obtain the conduction simulation model includes:

[0069] In the digital twin model, the corresponding density, specific heat capacity, and thermal conductivity are set for each component of the power generation equipment, and the corresponding boundary conditions are set for the entire digital twin model. The finite element method (FEM) is used to simulate the heat conduction of the digital twin model using ANSYS simulation software, and the time step and number of iterations of the heat conduction simulation are set;

[0070] The digital twin model at this time is marked as a conduction simulation model. The conduction simulation model can obtain the temperature distribution of the power generation equipment and render the obtained temperature distribution as the corresponding color temperature. The conduction simulation model can also simulate the heat conduction conditions at a single heat source in isolation to obtain its temperature conduction conditions that change with increasing conduction distance.

[0071] It should be further explained that, in a specific implementation process, the process of constructing the color temperature conduction area of ​​each color temperature detection point and obtaining the color temperature conduction parameters of each color temperature detection point includes:

[0072] The first adjusted color temperature of a single color temperature detection point is uploaded to the conduction simulation model for synchronization, and the temperature conduction condition of the color temperature detection point at its corresponding position is simulated by using the conduction simulation model, and the temperature conduction condition is represented by the color temperature change value at different conduction distances, the transmission distance refers to the shortest distance between any point and the color temperature detection point, and the color temperature change value refers to the difference between the color temperature of any point and the color temperature of the color temperature detection point;

[0073] In the conduction simulation model, a color temperature conduction area of ​​the color temperature detection point is constructed with the color temperature detection point as the center, wherein the color temperature change value on the boundary of the color temperature conduction area is equal to the color temperature of the color temperature monitoring point, indicating that the temperature conduction range of the color temperature detection point ends here;

[0074] The same method is adopted to respectively obtain the color temperature conduction area of ​​each color temperature detection point and its corresponding temperature conduction situation, construct the color temperature conduction area of ​​each color temperature detection point based on the second detection image, and obtain the intersection area between different color temperature conduction areas, wherein the intersection area is used to indicate that the area is simultaneously affected by the temperature conduction of multiple color temperature detection points;

[0075] Taking any color temperature detection point in the intersection area as an example, the color temperature change value of the color temperature conduction area generating the intersection area at the color temperature detection point is marked as B i , wherein i=1, 2, ..., n, n represents the number of color temperature conduction zones that generate the intersection area, and the color temperature conduction parameter is each color temperature change value at the color temperature detection point.

[0076] It should be further explained that, in a specific implementation process, the process of adjusting the second detection image according to the color temperature transmission parameter and obtaining the third detection image includes:

[0077] The color temperature detection points obtained by the present invention are the heat source points of the power generation equipment, and the other pixel points in the second detection image are located in the color temperature conduction area of ​​each color temperature detection point and different intersection areas, and the color temperature conduction parameters of each pixel point are obtained respectively;

[0078] According to the color temperature conductivity parameter of the pixel point, the first color temperature W corresponding to the pixel point is adjusted y Adjust to obtain the second adjusted color temperature W e ;

[0079]

[0080] Among them, W yi The first adjusted color temperature of the color temperature detection point corresponding to the i-th color temperature change value is obtained, and the second adjusted color temperature of each pixel in the second detection image is obtained respectively. A third detection image is obtained according to each pixel under the second adjusted color temperature after adjustment. The third detection image has the same area as the second detection image, but has a different color temperature.

[0081] It should be further explained that, in a specific implementation process, the process of setting the color temperature alarm threshold, judging whether there is an abnormal color temperature point in the third detection image and providing feedback includes:

[0082] Since different power generation equipment has different tasks in power generation, its normal working temperature is also different. Corresponding color temperature alarm thresholds are set for different power generation equipment. The color temperature alarm thresholds are all higher than the color temperature detection threshold.

[0083] The color temperature of each pixel in the third detection image is compared with its corresponding color temperature alarm threshold. If there is a pixel whose color temperature is greater than or equal to the color temperature alarm threshold, it is judged that there is an abnormal color temperature point, which means that the position of the corresponding pixel of the power generation equipment has a high temperature. An alarm signal is generated and fed back. If not, no other operation is performed on it.

[0084] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A safety fault detection system for power generation enterprises based on image recognition, characterized in that: Includes the following modules: The equipment twin module is used to collect equipment information of power generation equipment and build a digital twin model of the power generation equipment based on the equipment information; An infrared detection module is used to set infrared detection points, obtain an initial detection image of the power generation equipment, obtain a first detection image and several color temperature detection points in the initial detection image, obtain radiation parameters of the color temperature detection points in combination with the digital twin model, adjust the first detection image according to the radiation parameters, and obtain a second detection image; A conduction simulation module is used to simulate the heat conduction of the power generation equipment in the digital twin model to obtain a conduction simulation model, construct a color temperature conduction area of ​​each color temperature detection point, and obtain the color temperature conduction parameters of each color temperature detection point, adjust the second detection image according to the color temperature conduction parameters, and obtain a third detection image; A safety detection module is used to set a color temperature alarm threshold, determine whether there is an abnormal color temperature point in the third detection image, and provide feedback; The process of building a digital twin model of a power generation facility includes: The equipment information refers to the physical structural parameters of each power generation equipment, including the external shape and internal structure size parameters of the power generation equipment; Use digital twin technology to build digital twin models of each power generation equipment based on equipment information, and connect the digital twin models of different power generation equipment based on the connection relationship between the power generation equipment; The process of obtaining the initial detection image, the first detection image, and the color temperature detection point includes: A plurality of infrared detection points are respectively set in the working scene of the power generation equipment, and the surface temperature of the power generation equipment is detected through the infrared detection points and a corresponding initial detection image is generated; A color temperature detection threshold is set, and a local area formed by pixels in the initial detection image whose color temperature is greater than or equal to the color temperature detection threshold is used as a first detection image; An initial color temperature detection point is obtained in the first detection image, and the probability value of each other pixel point being selected as the next color temperature detection point is obtained. The pixel point corresponding to the maximum probability value is selected as the next color temperature detection point, and this step is repeated until m color temperature detection points are selected.

2. The power generation enterprise safety fault detection system based on image recognition according to claim 1 is characterized in that: The process of acquiring radiation parameters, adjusting the first detection image, and obtaining the second detection image includes: Upload the initial detection image to the digital twin model of the power generation equipment for synchronization, and obtain the first detection image and the corresponding position of its color temperature detection point in the digital twin model; The angle between the line connecting the color temperature detection point and the infrared detection point and the horizontal line is taken as the detection angle S a , obtain the surface roughness S of the color temperature detection point b and material property value S c , the radiation parameters include detection angle, surface roughness, and material property values; Get the initial color temperature W of the color temperature detection point c , the initial color temperature of the color temperature detection point is adjusted using the radiation parameter to obtain the first adjusted color temperature W y ; The radiation parameters of each pixel point in the first detection image are obtained, and the initial color temperature thereof is adjusted to obtain the corresponding first adjusted color temperature, and the second detection image is obtained according to each pixel point under the adjusted first adjusted color temperature.

3. The power generation enterprise safety fault detection system based on image recognition according to claim 2 is characterized in that: The process of obtaining a conduction simulation model includes: In the digital twin model, the corresponding density, specific heat capacity and thermal conductivity are set for each component of the power generation equipment, and the corresponding boundary conditions are set for the entire digital twin model. The finite element method is used to simulate the heat conduction of the digital twin model using ANSYS simulation software to obtain a conduction simulation model.

4. The power generation enterprise safety fault detection system based on image recognition according to claim 3 is characterized in that: The process of constructing the color temperature transmission area and obtaining the color temperature transmission parameters includes: The first adjusted color temperature of the color temperature detection point is uploaded to the conduction simulation model for synchronization, and the temperature conduction condition of the color temperature detection point at its corresponding position is simulated by the conduction simulation model, and the temperature conduction condition is represented by the color temperature change value at different conduction distances; In the conduction simulation model, a color temperature conduction area of ​​the color temperature detection point is constructed with the color temperature detection point as the center, and the color temperature change value on the boundary of the color temperature conduction area is equal to the color temperature of its color temperature monitoring point; Constructing a color temperature conduction area of ​​each color temperature detection point based on the second detection image, and obtaining an intersection area between different color temperature conduction areas; A color temperature detection point in the intersection area is obtained, and a plurality of color temperature change values ​​of a plurality of color temperature conduction zones in the intersection area at the color temperature detection point are used as color temperature conduction parameters.

5. The power generation enterprise safety fault detection system based on image recognition according to claim 4 is characterized in that: The process of adjusting the second detection image and obtaining the third detection image includes: According to each color temperature conduction area and the intersection area in the second detection image, a color temperature conduction parameter of each pixel point is obtained; According to the color temperature transmission parameter of the pixel point, the color temperature W is first adjusted. y Adjust to obtain the second adjusted color temperature W e ; Among them, W yi is the i-th color temperature transmission parameter B i The first adjusted color temperature of the corresponding color temperature detection point, i=1, 2, ..., n, n represents the number of color temperature conduction areas that produce its intersection area, and the third detection image is obtained according to each pixel point under the second adjusted color temperature after adjustment.

6. The power generation enterprise safety fault detection system based on image recognition according to claim 5 is characterized in that: The process of judging whether there is an abnormal color temperature point in the third detection image and providing feedback includes: Corresponding color temperature alarm thresholds are set for different power generation equipment, the color temperature of each pixel in the third detection image is compared with its corresponding color temperature alarm threshold, the pixel points that are greater than or equal to the color temperature alarm threshold are judged as abnormal color temperature points, and an alarm signal is generated and fed back.

Images