Electric fusion welding evaluation method, device and equipment based on thermal imaging and storage medium
Through the electrofusion welding evaluation method based on thermal imaging, the welding process is captured in real time and the deep learning and heat transfer model is used for detection and analysis, the problem that traditional detection methods cannot achieve non-destructive online detection is solved, and efficient and accurate welding quality evaluation is achieved.
Patent Information
- Application Number
- CN202411500969.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-25
- Publication Date
- 2025-03-25
AI Technical Summary
The traditional PE pipeline electrofusion welding quality detection method has the problem of inability to achieve non-destructive online inspection, inefficiency and comprehensive inspection, and it is difficult to meet actual needs.
The electrofusion welding evaluation method based on thermal imaging is adopted, and the welding process is captured in real time through infrared thermal imagers. U-Net is used to separate the welding areas. The CNN model detects temperature abnormalities, and the deep thermal characteristics are extracted based on heat transfer neural network model, the infrared image is reconstructed and the welding completion degree is judged.
Real-time, accurate and non-destructive online inspection of the welding process is achieved, the efficiency and comprehensiveness of welding quality evaluation is improved, and the stability and consistency of welding quality is ensured.
Smart Images
Figure CN119693293A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of PE pipe electric fusion welding quality assessment, and in particular to an electric fusion welding assessment method, device, equipment and computer-readable storage medium based on thermal imaging. Background Art
[0002] PE pipe electric fusion welding is a commonly used PE pipe connection method, and its welding quality is crucial to the safe operation of the pipeline. However, traditional welding quality inspection methods have many limitations. For example, destructive tests are not online, and it is impossible to perform process inspections on all welding parts; manual inspections rely on the experience and skills of inspectors and are inefficient. As a result, traditional inspection methods cannot achieve non-destructive online inspections, and their comprehensiveness and efficiency are difficult to meet actual needs.
[0003] With the continuous development of science and technology, there are higher requirements for the quality inspection of PE pipe electric fusion welding, which requires a more efficient, accurate and non-destructive online inspection method. At present, infrared thermal imaging technology has shown great potential in industrial inspection. How to make full use of infrared thermal imaging technology to achieve rapid and accurate evaluation of welding quality is a technical problem that needs to be solved urgently. Summary of the invention
[0004] The main purpose of the present invention is to provide an electric fusion welding evaluation method, device, equipment and computer-readable storage medium based on thermal imaging, aiming to solve the technical problems raised by the above-mentioned background technology.
[0005] To achieve the above object, the present invention provides an electric fusion welding quality assessment method based on thermal imaging, the electric fusion welding quality assessment method based on thermal imaging comprising the following steps:
[0006] The electric fusion welding process of PE pipe fittings is photographed in real time by infrared thermal imager, and the image data is collected periodically;
[0007] Based on U-Net, the infrared image of the welding area is separated according to the edge of the sleeve;
[0008] Input the welding infrared image into the trained CNN model and output the image with abnormal temperature areas;
[0009] The welding conditions and infrared images are input into the neural network model trained by the database established based on heat transfer to extract deep thermal features;
[0010] Complete deep infrared image reconstruction and welding completion judgment based on deep thermal characteristics;
[0011] Make sure that the deep temperature reaches the welding target temperature, end welding, and turn off the welding machine.
[0012] The method of separating the infrared image of the welding area according to the edge of the sleeve based on U-Net includes:
[0013] The infrared image I is input into the U-Net model, the infrared image I is preprocessed, the preprocessed image is processed, and the separated infrared image of the welding area is output.
[0014] Input the infrared image I into the U-Net model, preprocess the infrared image I, process the preprocessed image, and output the separated infrared image of the welding area, including:
[0015] Input infrared image I,
[0016] Perform Gaussian blur denoising on the infrared image I:
[0017] I blur =G(I,σ)
[0018] Perform normalization:
[0019]
[0020] The operations of the encoder layer are:
[0021] E l =ReLU(Conv2D(E l-1 ,W l )+b l )
[0022] Where E0 = I, E l is the feature map of layer l, W l and b l They are the convolution kernel and bias of the lth layer respectively;
[0023] The operation of the decoder layer is:
[0024] D l =ReLU(Conv2D(UpSampling2D(D l+1 ),W l ')+b l ')
[0025] Among them, D L-1 =Z,D l is the feature map of layer l, W l ' and b l ' are the convolution kernel and bias of the lth layer respectively;
[0026] In each layer of the decoder, the feature map of the corresponding layer of the encoder is concatenated with the feature map of the decoder through a skip connection:
[0027] D l=Concat(E L-l-1 ,D l )
[0028] The last layer outputs the segmentation result:
[0029]
[0030] Loss function:
[0031]
[0032] Where N is the number of samples, y i is the true label, is the model prediction.
[0033] The welding infrared image is input into the trained CNN model, and the image with abnormal temperature area is output, including:
[0034] The welding infrared image is passed into the convolution layer to extract the local features of the image. The mathematical expression is:
[0035]
[0036] II is the input image, K is the convolution kernel, b is the bias term, and F ij is the value of the output feature map at position (i, j)
[0037] The output of the convolutional layer is processed by the nonlinear activation function Relu:
[0038] ReLU(x)=max(0,x)
[0039] Then enter the pooling layer for downsampling to reduce the size of the feature map. Average pooling is used here, and the mathematical formula is as follows:
[0040]
[0041] Among them, F ij Represents the value of the feature map after pooling at the position, h p Indicates the height of the pooling window, w p represents the width of the pooling window; F is the input feature map, i and j represent the index of the position where the pooling window slides on the feature map, s p represents the step size of the pooling window sliding, m and n represent the index of the local area in the pooling window;
[0042] After convolution and pooling, the data is passed to the fully connected layer for classification. The fully connected layer is:
[0043] y=Wx+b
[0044] After CNN classification is completed, abnormal areas are highlighted through class activation mapping:
[0045]
[0046] Among them, F k (x,y) is the kth feature map of the last convolutional layer, w k is the weight of the corresponding category of the fully connected layer;
[0047] After the temperature field is detected by CNN, the image of welding abnormality and the abnormal area are output.
[0048] The neural network model trained by the database established based on heat transfer includes:
[0049] According to the heat transfer theory, combined with the physical properties of PE material and resistance wire, the following transient heat conduction equation is used to solve the temperature field of PE pipe electric fusion welding joint:
[0050]
[0051] In the formula, λ is the thermal conductivity, c is the specific heat capacity, ρ is the density, and γ is the position of the resistance wire;
[0052] The contact surface between the electric fusion pipe and the air is regarded as the thermal convection boundary condition, and the resulting thermal boundary condition is:
[0053]
[0054] In the formula, α1 represents the convection heat transfer coefficient of the inner surface of the electric fusion pipe fitting, and α2 represents the convection heat transfer coefficient of the outer surface of the electric fusion pipe fitting;
[0055] The calculation expression of heat Q generated by the resistance wire is as follows:
[0056]
[0057] Where R is the calibrated resistance, β is the temperature coefficient of resistance, T' is the temperature when the calibrated resistance value R is reached, and U is the welding voltage.
[0058] If the maximum welding temperature of the weld reaches 220°C, turn off the welder.
[0059] In addition, to achieve the above-mentioned purpose, the present invention also provides an electric fusion welding evaluation device based on thermal imaging, and the electric fusion welding evaluation device based on thermal imaging comprises:
[0060] The acquisition module is used to take real-time photos of the PE pipe electric fusion welding process through an infrared thermal imager and periodically collect the captured image data;
[0061] A separation module is used to separate the infrared image of the welding area according to the edge of the sleeve based on U-Net;
[0062] The anomaly assessment module is used to input the welding infrared image into the trained CNN model and output the image with the temperature abnormality area;
[0063] An analysis module is used to input welding conditions and infrared images into a neural network model trained by a database based on heat transfer to extract deep thermal features;
[0064] A judgment module is used to complete deep infrared image reconstruction and welding completion judgment based on deep thermal characteristics;
[0065] The end module is used to determine when the deep temperature reaches the welding target temperature, end welding, and shut down the welder.
[0066] In addition, to achieve the above-mentioned purpose, the present invention also provides an electric fusion welding evaluation device based on thermal imaging, which includes an electric fusion welding machine, an infrared thermal imager, a processor, a memory, and an electric fusion welding evaluation program based on thermal imaging stored in the memory and executable by the processor, wherein when the electric fusion welding evaluation program based on thermal imaging is executed by the processor, the steps of the electric fusion welding evaluation method based on thermal imaging are implemented.
[0067] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a thermal imaging-based electric fusion welding evaluation program is stored, wherein when the thermal imaging-based electric fusion welding evaluation program is executed by a processor, the steps of the thermal imaging-based electric fusion welding evaluation method are implemented.
[0068] The present invention provides an electric fusion welding evaluation method based on thermal imaging, wherein the method uses an infrared thermal imager to shoot the electric fusion welding process of PE pipe fittings in real time, and periodically collects the image data of the shooting; based on U-Net, the infrared image of the welding area is separated according to the edge of the sleeve; the welding infrared image is input into a trained CNN model, and an image of the temperature abnormality area is output; the welding condition and the infrared image are input into a neural network model trained by a database established based on heat transfer to extract deep thermal features; deep infrared image reconstruction and welding completion judgment are completed according to the deep thermal features; it is determined that the deep temperature reaches the welding standard temperature, the welding is completed, and the welding machine is turned off. By using an infrared thermal imager to shoot the electric fusion welding process of PE pipe fittings in real time and periodically collect image data, the temperature change during the welding process can be grasped in real time, thereby realizing accurate control of the welding process. When it is determined that the deep temperature reaches the welding standard temperature (such as 220°C), the welding machine is turned off in time to ensure the stability and consistency of the welding quality. Accurate segmentation and anomaly detection: UNet is used to separate the infrared image of the welding area according to the edge of the sleeve, which improves the accuracy of welding area positioning. Inputting the welding infrared image into the trained CNN model can quickly output images of areas with abnormal temperatures, detect welding quality problems in a timely manner, and provide a basis for subsequent repairs and improvements. Deep thermal feature analysis and welding completion judgment: The welding conditions and infrared images are input into the neural network model trained by the database established based on heat transfer to obtain deep thermal features, and then complete deep infrared image reconstruction and welding completion judgment. This helps to fully understand the heat distribution during the welding process, accurately evaluate the welding quality, and ensure the safety and reliability of the pipeline. Non-destructive testing and cost reduction: The entire detection process is non-destructive and will not cause damage to the pipeline, reducing the detection cost and pipeline maintenance cost, avoiding pipeline damage caused by destructive testing, and has good economic and environmental benefits. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] Figure 1 It is a hardware schematic diagram of the electric fusion welding evaluation equipment based on thermal imaging of the present invention;
[0070] Figure 2 It is a schematic flow chart of the electric fusion welding evaluation method based on thermal imaging of the present invention;
[0071] Figure 3 It is a schematic diagram of the overall process of the electric fusion welding evaluation method based on thermal imaging of the present invention;
[0072] Figure 4 This is a schematic diagram of the image classification network and temperature difference prediction network training of the present invention;
[0073] Figure 5 It is a schematic diagram of the overall process simulation of the present invention.
[0074] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0075] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.
[0076] The electric fusion welding evaluation method based on thermal imaging involved in the embodiment of the present invention is mainly applied to an electric fusion welding evaluation device based on thermal imaging, which can be a device with display and processing functions such as a PC, a portable computer, or a mobile terminal.
[0077] Reference Figure 1 , Figure 1 The figure is a schematic diagram of the hardware structure of the electric fusion welding evaluation device based on thermal imaging involved in the embodiment of the present invention. In the embodiment of the present invention, the electric fusion welding evaluation device based on thermal imaging may include a processor (such as a CPU), a communication bus, a user interface, a network interface, and a memory. Among them, the communication bus is used to realize the connection and communication between these components; the user interface may include a display (Display), an input unit such as a keyboard (Keyboard); the network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface); the memory may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory, and the memory may optionally be a storage device independent of the aforementioned processor.
[0078] Those skilled in the art will understand that Figure 1 The hardware structure shown in the figure does not constitute a limitation on the electric fusion welding evaluation device based on thermal imaging, and may include more or less components than those shown in the figure, or combine certain components, or arrange the components differently.
[0079] Continue to refer to Figure 1 , Figure 1 The memory as a computer-readable storage medium may include an operating system, a network communication module, and a thermal imaging-based electrofusion welding evaluation program.
[0080] exist Figure 1 In the embodiment, the network communication module is mainly used to connect to the server and perform data communication with the server; and the processor can call the electric fusion welding evaluation program based on thermal imaging stored in the memory, and execute the electric fusion welding evaluation method based on thermal imaging provided in the embodiment of the present invention.
[0081] An embodiment of the present invention provides an electric fusion welding evaluation method based on thermal imaging.
[0082] Reference Figure 2-Figure 5, which is a schematic diagram related to the electric fusion welding evaluation method based on thermal imaging of the present invention.
[0083] The present invention provides an electric fusion welding evaluation method based on thermal imaging, and the electric fusion welding evaluation method based on thermal imaging comprises the following steps:
[0084] The electric fusion welding process of PE pipe fittings is photographed in real time by infrared thermal imager, and the image data is collected periodically;
[0085] Based on U-Net, the infrared image of the welding area is separated according to the edge of the sleeve;
[0086] Input the welding infrared image into the trained CNN model and output the image with abnormal temperature areas;
[0087] The welding conditions and infrared images are input into the neural network model trained by the database established based on heat transfer to extract deep thermal features;
[0088] Complete deep infrared image reconstruction and welding completion judgment based on deep thermal characteristics;
[0089] Make sure that the deep temperature reaches the welding target temperature, end welding, and turn off the welding machine.
[0090] In this embodiment, the present invention provides an electric fusion welding evaluation method based on thermal imaging, the method uses an infrared thermal imager to shoot the electric fusion welding process of PE pipe fittings in real time, and periodically collects the image data taken; based on U-Net, the infrared image of the welding area is separated according to the edge of the sleeve; the welding infrared image is input into the trained CNN model, and the image of the temperature abnormality area is output; the welding condition and the infrared image are input into the neural network model trained by the database established based on heat transfer to extract the deep thermal features; the deep infrared image reconstruction and welding completion judgment are completed according to the deep thermal features; it is determined that the deep temperature reaches the welding standard temperature, the welding is completed, and the welder is turned off. By shooting the electric fusion welding process of PE pipes in real time and periodically collecting image data with an infrared thermal imager, the temperature change during the welding process can be grasped in real time, thereby realizing accurate control of the welding process. When it is determined that the deep temperature reaches the welding standard temperature (such as 220°C), the welder is turned off in time to ensure the stability and consistency of the welding quality. Accurate segmentation and anomaly detection: U-Net is used to separate the infrared image of the welding area according to the edge of the sleeve, which improves the accuracy of the positioning of the welding area. Inputting the welding infrared image into the trained CNN model can quickly output images of areas with abnormal temperatures, promptly detect welding quality problems, and provide a basis for subsequent repairs and improvements. Deep thermal feature analysis and welding completion judgment: The welding conditions and infrared images are input into the neural network model trained by the database established based on heat transfer to obtain deep thermal features, and then complete deep infrared image reconstruction and welding completion judgment. This helps to fully understand the heat distribution during the welding process, accurately evaluate the welding quality, and ensure the safety and reliability of the pipeline. Non-destructive testing and cost reduction: The entire detection process is non-destructive and will not cause damage to the pipeline, reducing the detection cost and pipeline maintenance cost, avoiding pipeline waste due to destructive testing, and has good economic and environmental benefits.
[0091] Temperature field reconstruction based on heat transfer, temperature difference prediction proxy model and infrared thermal imaging:
[0092] In actual testing, an infrared thermal imager is used to obtain the surface temperature.
[0093] The actual welding conditions are input into the temperature difference prediction model to obtain the temperature difference ΔT predicted (x,r,t).
[0094] According to the predicted temperature difference ΔT predicted (x,r,t), calculate the temperature of the welding area
[0095] Welding completion judgment based on welds:
[0096] If the maximum welding temperature of the weld reaches 220°C, turn off the welder.
[0097] The method of separating the infrared image of the welding area according to the edge of the sleeve based on U-Net includes:
[0098] The infrared image I is input into the U-Net model, the infrared image I is preprocessed, the preprocessed image is processed, and the separated infrared image of the welding area is output.
[0099] Input the infrared image I into the U-Net model, preprocess the infrared image I, process the preprocessed image, and output the separated infrared image of the welding area, including:
[0100] Input infrared image I,
[0101] Perform Gaussian blur denoising on the infrared image I:
[0102] I blur =G(I,σ)
[0103] Perform normalization:
[0104]
[0105] The operations of the encoder layer are:
[0106] E l =ReLU(Conv2D(E l-1 ,W l )+b l )
[0107] Where E0 = I, E l is the feature map of layer l, W l and b l They are the convolution kernel and bias of the lth layer respectively;
[0108] The operation of the decoder layer is:
[0109] D l =ReLU(Conv2D(UpSampling2D(D l+1 ),W l ')+b l ')
[0110] Among them, D L-1 =Z,D l is the feature map of layer l, W l ' and b l ' are the convolution kernel and bias of the lth layer respectively;
[0111] In each layer of the decoder, the feature map of the corresponding layer of the encoder is concatenated with the feature map of the decoder through a skip connection:
[0112] D l =Concat(E L-l-1 ,D l )
[0113] The last layer outputs the segmentation result:
[0114]
[0115] Loss function:
[0116]
[0117] Where N is the number of samples, y i is the true label, is the model prediction.
[0118] The welding infrared image is input into the trained CNN model, and the image with abnormal temperature area is output, including:
[0119] The welding infrared image is passed into the convolution layer to extract the local features of the image. The mathematical expression is:
[0120]
[0121] II is the input image, K is the convolution kernel, b is the bias term, and F ij is the value of the output feature map at position (i, j).
[0122] The output of the convolutional layer is processed by the nonlinear activation function Relu:
[0123] ReLU(x)=max(0,x)
[0124] Then enter the pooling layer for downsampling to reduce the size of the feature map. Average pooling is used here, and the mathematical formula is as follows:
[0125]
[0126] Among them, F ij Represents the value of the feature map after pooling at the position, h p Indicates the height of the pooling window, w p represents the width of the pooling window; F is the input feature map, i and j represent the index of the position where the pooling window slides on the feature map, s p represents the step size of the pooling window sliding, m and n represent the index of the local area in the pooling window;
[0127] After convolution and pooling, the data is passed to the fully connected layer for classification. The fully connected layer is:
[0128] y=Wx+b
[0129] After CNN classification is completed, abnormal areas are highlighted through class activation mapping:
[0130]
[0131] Among them, F k (x,y) is the kth feature map of the last convolutional layer, w k is the weight of the corresponding category of the fully connected layer;
[0132] After the temperature field is detected by CNN, the image of welding abnormality and the abnormal area are output.
[0133] The neural network model trained by the database established based on heat transfer includes:
[0134] According to the heat transfer theory, combined with the physical properties of PE material and resistance wire, the following transient heat conduction equation is used to solve the temperature field of PE pipe electric fusion welding joint:
[0135]
[0136] In the formula, λ is the thermal conductivity, c is the specific heat capacity, ρ is the density, and γ is the position of the resistance wire;
[0137] The contact surface between the electric fusion pipe and the air is regarded as the thermal convection boundary condition, and the resulting thermal boundary condition is:
[0138]
[0139] In the formula, α1 represents the convection heat transfer coefficient of the inner surface of the electric fusion pipe fitting, and α2 represents the convection heat transfer coefficient of the outer surface of the electric fusion pipe fitting;
[0140] The calculation expression of heat Q generated by the resistance wire is as follows:
[0141]
[0142] Where R is the calibrated resistance, β is the temperature coefficient of resistance, T' is the temperature when the calibrated resistance value R is reached, and U is the welding voltage.
[0143] If the maximum welding temperature of the weld reaches 220°C, turn off the welder.
[0144] In addition, to achieve the above-mentioned purpose, the present invention also provides an electric fusion welding evaluation device based on thermal imaging, and the electric fusion welding evaluation device based on thermal imaging comprises:
[0145] The acquisition module is used to take real-time photos of the PE pipe electric fusion welding process through an infrared thermal imager and periodically collect the captured image data;
[0146] A separation module is used to separate the infrared image of the welding area according to the edge of the sleeve based on U-Net;
[0147] The anomaly assessment module is used to input the welding infrared image into the trained CNN model and output the image with the temperature abnormality area;
[0148] An analysis module is used to input welding conditions and infrared images into a neural network model trained by a database based on heat transfer to extract deep thermal features;
[0149] A judgment module is used to complete deep infrared image reconstruction and welding completion judgment based on deep thermal characteristics;
[0150] The end module is used to determine when the deep temperature reaches the welding target temperature, end welding, and shut down the welder.
[0151] In addition, to achieve the above-mentioned purpose, the present invention also provides an electric fusion welding evaluation device based on thermal imaging, which includes an electric fusion welding machine, an infrared thermal imager, a processor, a memory, and an electric fusion welding evaluation program based on thermal imaging stored in the memory and executable by the processor, wherein when the electric fusion welding evaluation program based on thermal imaging is executed by the processor, the steps of the electric fusion welding evaluation method based on thermal imaging are implemented.
[0152] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which is stored an electric fusion welding evaluation program based on thermal imaging, wherein when the electric fusion welding evaluation program based on thermal imaging is executed by a processor, the steps of the electric fusion welding evaluation method based on thermal imaging are implemented.
[0153] The method implemented when the electric fusion welding evaluation program based on thermal imaging is executed can refer to the various embodiments of the electric fusion welding evaluation method based on thermal imaging of the present invention, which will not be described in detail here.
[0154] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.
[0155] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.
[0156] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.
[0157] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.
[0158] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A method for evaluating electric fusion welding based on thermal imaging, characterized in that: The electric fusion welding evaluation method based on thermal imaging comprises the following steps: The electric fusion welding process of PE pipe fittings is photographed in real time by infrared thermal imager, and the image data is collected periodically; Based on U-Net, the infrared image of the welding area is separated according to the edge of the sleeve; Input the welding infrared image into the trained CNN model and output the image with abnormal temperature areas; The welding conditions and infrared images are input into the neural network model trained by the database established based on heat transfer to extract deep thermal features; Complete deep infrared image reconstruction and welding completion judgment based on deep thermal characteristics; Make sure that the deep temperature reaches the welding target temperature, end welding, and turn off the welding machine.
2. The electric fusion welding evaluation method based on thermal imaging according to claim 1, characterized in that: The method of separating the infrared image of the welding area according to the edge of the sleeve based on U-Net includes: The infrared image I is input into the U-Net model, the infrared image I is preprocessed, the preprocessed image is processed, and the separated infrared image of the welding area is output.
3. The electric fusion welding evaluation method based on thermal imaging according to claim 2, characterized in that: The input infrared image I is input into the U-Net model, the infrared image I is preprocessed, the preprocessed image is processed, and the separated infrared image of the welding area is output, including: Input infrared image I, Perform Gaussian blur denoising on the infrared image I: I blur =G(I,σ) Perform normalization: The operations of the encoder layer are: It is l =ReLU(Conv2D(E l-1 ,W l )+b l ) Where E0 = I, E l is the feature map of layer l, W l and b l They are the convolution kernel and bias of the lth layer respectively; The operation of the decoder layer is: D l =ReLU(Conv2D(UpSampling2D(D l+1 ),W l ')+b l ') Among them, D L-1 =Z,D l is the feature map of layer l, W l ' and b l ' are the convolution kernel and bias of the lth layer respectively; In each layer of the decoder, the feature map of the corresponding layer of the encoder is concatenated with the feature map of the decoder through a skip connection: D l =Concat(E L-l-1 ,D l ) The last layer outputs the segmentation result: Loss function: Where N is the number of samples, y i is the true label, is the model prediction.
4. The electric fusion welding evaluation method based on thermal imaging according to claim 1, characterized in that: The welding infrared image is input into the trained CNN model, and the image with abnormal temperature area is output, including: The welding infrared image is passed into the convolution layer to extract the local features of the image. The mathematical expression is: II is the input image, K is the convolution kernel, b is the bias term, and F ij is the value of the output feature map at position (i, j) The output of the convolutional layer is processed by the nonlinear activation function Relu: ReLU(x)=max(0,x) Then enter the pooling layer for sampling to reduce the size of the feature map. Average pooling is used here. The mathematical formula is as follows: Among them, F ij Represents the value of the feature map after pooling at the position, h p Indicates the height of the pooling window, w p represents the width of the pooling window; F is the input feature map, i and j represent the index of the position where the pooling window slides on the feature map, s p represents the step size of the pooling window sliding, m and n represent the index of the local area in the pooling window; After convolution and pooling, the data is passed to the fully connected layer for classification. The fully connected layer is: y=Wx+b After CNN classification is completed, abnormal areas are highlighted through class activation mapping: Among them, F k (x,y) is the kth feature map of the last convolutional layer, w k is the weight of the corresponding category of the fully connected layer; After the temperature field is detected by CNN, the image of welding abnormality and the abnormal area are output.
5. The method for evaluating electric fusion welding based on thermal imaging according to claim 1, characterized in that: The neural network model trained by the database established based on heat transfer includes: According to the heat transfer theory, combined with the physical properties of PE material and resistance wire, the following transient heat conduction equation is used to solve the temperature field of PE pipe electric fusion welding joint: In the formula, λ is the thermal conductivity, c is the specific heat capacity, ρ is the density, and γ is the position of the resistance wire; The contact surface between the electric fusion pipe and the air is regarded as the thermal convection boundary condition, and the resulting thermal boundary condition is: In the formula, α1 represents the convection heat transfer coefficient of the inner surface of the electric fusion pipe fitting, and α2 represents the convection heat transfer coefficient of the outer surface of the electric fusion pipe fitting; The calculation expression of heat Q generated by the resistance wire is as follows: Where R is the calibrated resistance, β is the temperature coefficient of resistance, T' is the temperature when the calibrated resistance value R is reached, and U is the welding voltage.
6. The method for evaluating electric fusion welding based on thermal imaging according to claim 1, characterized in that: If the maximum temperature of the welding zone reaches 220℃, welding is completed.
7. An electric fusion welding evaluation device based on thermal imaging, characterized in that: The electric fusion welding evaluation device based on thermal imaging comprises: The acquisition module is used to take real-time photos of the PE pipe electric fusion welding process through an infrared thermal imager and periodically collect the captured image data; A separation module is used to separate the infrared image of the welding area according to the edge of the sleeve based on U-Net; The anomaly assessment module is used to input the welding infrared image into the trained CNN model and output the image with the temperature abnormality area; An analysis module is used to input welding conditions and infrared images into a neural network model trained by a database based on heat transfer to extract deep thermal features; A judgment module is used to complete deep infrared image reconstruction and welding completion judgment based on deep thermal characteristics; The end module is used to determine when the deep temperature reaches the welding target temperature, end welding, and shut down the welder.
8. An electric fusion welding evaluation device based on thermal imaging, characterized in that: The electric fusion welding evaluation device based on thermal imaging includes an electric fusion welding machine, an infrared thermal imager, a processor, a memory, and an electric fusion welding evaluation program based on thermal imaging stored in the memory and executable by the processor, wherein when the electric fusion welding evaluation program based on thermal imaging is executed by the processor, the steps of the electric fusion welding evaluation method based on thermal imaging as described in any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a thermal imaging-based electric fusion welding evaluation program, wherein when the thermal imaging-based electric fusion welding evaluation program is executed by a processor, the steps of the thermal imaging-based electric fusion welding evaluation method according to any one of claims 1 to 6 are implemented.