A stress measurement method, a stress measurement system, and a computer-readable medium.

Abstract

This application provides a stress measurement method, a stress measurement system, and a computer-readable medium, which can solve the technical problem of low accuracy in manual residual stress detection. The stress measurement method includes: acquiring image information; preprocessing the image information to obtain capability characterization information; using the capability characterization information to characterize the stress on the pipeline structure; dividing the pipeline into a first training set and a first test set based on the capability characterization information; establishing a first stress detection model based on the first training set; and optimizing the first stress detection model based on the first test set to obtain a second stress detection model; measuring actual stress information; and optimizing the second stress detection model based on the actual stress information to obtain a third stress detection model. By preprocessing the image information, learning through a neural network, establishing the first stress detection model, and optimizing the model to finally obtain the third stress detection model, the accuracy of residual stress detection is improved.

Description

Technical Field

[0001] This application relates to the field of pipeline safety technology, and in particular to a stress measurement method, a stress measurement system, and a computer-readable medium. Background Technology

[0002] Pipeline structures are widely used in industries such as petroleum, chemical, and power. As a fundamental structural element, the reliability of pipeline structures for safe operation is crucial. Stress acting on a pipeline structure can reduce its load-bearing capacity, thus affecting its safe use. To ensure the reliability of pipeline structures for safe operation, it is necessary to measure the stress within the pipeline structure for reliability assessment.

[0003] Currently, stress in pipeline structures is detected manually. For example, residual stress within the joints of the pipeline structure is manually checked.

[0004] However, the accuracy of manual detection of residual stress is low. Summary of the Invention

[0005] This application provides a stress measurement method, a stress measurement system, and a computer-readable medium, which can solve the technical problem of low accuracy in manual detection of residual stress.

[0006] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0007] In a first aspect, a stress measurement method is provided, comprising: acquiring image information; preprocessing the image information to obtain capability characterization information; using the capability characterization information to characterize the stress on the pipeline structure; dividing the pipeline into a first training set and a first test set based on the capability characterization information; establishing a first stress detection model based on the first training set; and optimizing the first stress detection model based on the first test set to obtain a second stress detection model; measuring actual stress information; and optimizing the second stress detection model based on the actual stress information to obtain a third stress detection model.

[0008] Based on the above description of the stress measurement method provided in the embodiments of this application, it can be seen that the stress measurement method includes preprocessing image information, learning through a neural network to establish a first stress detection model, and optimizing the model to improve the accuracy of residual stress detection.

[0009] Furthermore, it eliminates the need for manual modeling by staff, simplifying operation. It also eliminates the need for manual residual stress detection, and while ensuring the accuracy of residual stress detection, it presents the data to staff in a more intuitive manner.

[0010] In the feasible implementation of the first aspect, after performing the steps of measuring actual stress information and optimizing the second stress detection model based on the actual stress information, the stress measurement method further includes: obtaining the measurement result based on the third stress detection model; and uploading the measurement result to the blockchain storage.

[0011] This method uploads measurement data to the blockchain for storage, preventing data tampering, facilitating subsequent verification of data accuracy by staff, reducing energy consumption during blockchain data processing, and improving its data processing efficiency.

[0012] In the feasible implementation of the first aspect, the measurement results after each group of identifications are processed into blocks that meet the conditions. When joining the network, each node in the blockchain network generates a local public-private key pair as its own identifier in the network. When a node is waiting for its local role to become a candidate node, it broadcasts a leader application to other nodes in the network and sends it. If the verification is successful, the relevant verification data is returned to the candidate node. When the candidate node becomes the leader node, other nodes become follower nodes. Then the leader node confirms the follower nodes in the block header. At the same time, the follower node sends a confirmation application to the leader node. The leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become the leader node again until the dormant period ends. After receiving the confirmation information, the follower nodes add the newly generated blocks to the chain and return the candidate identity.

[0013] In the feasible implementation of the first aspect, there are multiple sets of image information. When performing the step of preprocessing the image information to obtain capability characterization information, the stress measurement method further includes: processing the multiple sets of image information into the same format to obtain first information; converting the first information into second information through normalization; the value of the second information is greater than zero and less than one; obtaining the variance coefficient based on the second information; and obtaining the capability characterization information based on the variance coefficient.

[0014] In a feasible implementation of the first aspect, when performing the steps of establishing a first stress detection model based on a first training set and optimizing the first stress detection model based on a first test set to obtain a second stress detection model, the stress measurement method further includes: constructing a first convolutional neural network; and importing the first training set into the convolutional neural network to obtain the first stress detection model.

[0015] In a feasible implementation of the first aspect, when performing the steps of establishing a first stress detection model based on a first training set and optimizing the first stress detection model based on a first test set to obtain a second stress detection model, the stress measurement method further includes: inputting the first test set into the first stress detection model to obtain a first test result; if the first test result is less than a preset expected value, then repeating the step of importing the first training set into a convolutional neural network to obtain the first stress detection model.

[0016] In the feasible implementation of the first aspect, the image information includes the shape of the molten pool and the energy distribution; when performing the step of optimizing the second stress detection model based on the actual stress information to obtain the third stress detection model, the stress measurement method further includes: inputting the actual stress information into the second stress detection model; describing the first heat flux density distribution using a Gaussian function; constructing a double ellipsoidal heat source model based on the image information; obtaining the second heat flux density distribution based on the double ellipsoidal heat source model; inputting the first heat flux density distribution and the second heat flux density distribution into the second stress detection model, dividing it into a second training set and a second test set; standardizing the second training set to obtain training samples; importing the training samples into the second stress detection model, and training the second stress detection model using an iterative method to obtain the third stress detection model.

[0017] In the feasible implementation of the first aspect, the image information includes the shape of the molten pool and the energy distribution; after performing the step of optimizing the second stress detection model based on the actual stress information to obtain the third stress detection model, the stress measurement method further includes: inputting the second test set into the third stress detection model to obtain the residual stress measurement curve.

[0018] In the feasible implementation of the first aspect, after the step of acquiring image information and before the step of preprocessing the image information to obtain capability characterization information, the stress measurement method further includes: extracting the image information frame by frame to obtain a toroidal image; performing block processing according to the display ratio of the toroidal image to obtain an image block result; extracting the high-frequency components of the image block result through Fourier transform to obtain an image high-frequency result; smoothing the image high-frequency result through Gaussian filtering to obtain an image smoothing result; performing scale normalization processing on the toroidal image in the image smoothing result through an image pyramid to obtain an image normalization result; and processing the image normalization result through a backbone convolutional neural network. Feature extraction is performed on the toroidal image of the fruit to obtain image feature extraction results; the image feature extraction results are input into a bidirectional feature pyramid to obtain feature fusion results; the feature fusion results are classified and regressed to output detection boxes and categories; the coordinates of the detection boxes are generated; the toroidal image is enlarged and cropped to obtain a cropped image; simple negative samples in the cropped image are filtered to obtain a set of filtered images; the regions of the detection device in the set of filtered images are classified and regressed to obtain the first classification regression result; anchor boxes are generated in the cropped image, and the anchor boxes are classified and regressed to obtain the second classification regression result; based on the first and second classification results, target information in the cropped image is extracted through enlargement and cropping.

[0019] Secondly, embodiments of this application provide a stress measurement system, which includes: at least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method provided in the first aspect.

[0020] The stress measurement system performs the stress measurement method provided in the first aspect, preprocesses the image information, establishes a first stress detection model through neural network learning, and optimizes the model to improve the accuracy of residual stress detection.

[0021] Thirdly, embodiments of this application provide a computer-readable medium having computer program instructions stored thereon, which can be executed by a processor to implement the method provided in the first aspect.

[0022] The computer program instructions of the computer-readable medium can be executed to implement the method provided in the first aspect, preprocessing image information, learning through a neural network, establishing a first stress detection model, and optimizing the model to improve the accuracy of residual stress detection. Attached Figure Description

[0023] Figure 1 A schematic flowchart of one embodiment of the stress measurement method provided in this application;

[0024] Figure 2 A schematic flowchart of one embodiment of the stress measurement method provided in this application;

[0025] Figure 3 A schematic flowchart of one embodiment of the stress measurement method provided in this application;

[0026] Figure 4 A schematic flowchart of one embodiment of the stress measurement method provided in this application;

[0027] Figure 5 A schematic flowchart of one embodiment of the stress measurement method provided in this application;

[0028] Figure 6 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Detailed Implementation

[0029] The principles and features of this application are described below. The examples given are only for explaining this application and are not intended to limit the scope of this application.

[0030] This application provides a stress measurement method that can be used for stress measurement in pipeline structures. For example, it can be used to measure the residual stress in the welds of pipeline steel. There are various types of pipeline steel, such as X80.

[0031] In some embodiments, the pipeline structure may be an X80 ring.

[0032] Residual stress in welds refers to the residual stress generated within the joints of pipeline structures during the welding process due to uneven heating and cooling. The magnitude of this residual stress has a significant impact on the load-bearing capacity and safety reliability of the pipeline structure.

[0033] In particular, the presence of residual tensile stress, when superimposed with working stress, not only reduces the load-bearing capacity of the pipeline structure, but may even cause pipeline structure damage or premature failure.

[0034] Furthermore, the residual tensile stress on the inner surface of the pipeline, combined with the corrosive medium it transports, can cause stress corrosion cracking.

[0035] Therefore, residual stress is an important factor affecting the reliability assessment of pipeline safety.

[0036] The stress measurement method provided in this application can improve the accuracy of stress detection.

[0037] The following section uses the residual stress of the X80 circumferential weld as an example to illustrate the stress measurement method provided in the embodiments of this application in detail with reference to the accompanying drawings.

[0038] Figure 1 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Figure 1 As shown, in some embodiments, the stress measurement method includes the following steps:

[0039] S1, acquire image information.

[0040] In some embodiments, image information of the X80 ring is acquired using a detection device. The acquired image information needs to be recorded. In one implementation, the acquired image information can be stored in a storage unit within the stress measurement system.

[0041] In some embodiments, the detection device may be a sensor.

[0042] In some embodiments, image information of the X80 ring measurement location can be measured by setting a sensor at the measurement location on the X80 ring.

[0043] To ensure the accuracy of the measurement results, in some embodiments, multiple identical sensors and / or multiple different sensors can be deployed at the measurement location on the X80 ring. In one implementation, multiple sensors are deployed at circumferential intervals around the X80 ring. This results in multiple sets of acquired image information.

[0044] In some embodiments, the sensor may include a thermal sensor.

[0045] In some embodiments, the sensor may include an ultrasonic sensor.

[0046] In some embodiments, during the welding process, the inspection equipment acquires image information. This results in better data integrity and more accurate feedback.

[0047] S2, preprocesses the image information to obtain capability characterization information.

[0048] Figure 2 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Figure 2 As shown, in some embodiments, there are multiple sets of image information. When performing step S2, the stress measurement method further includes the following steps:

[0049] S21, process multiple sets of image information into the same format to obtain the first information.

[0050] In some embodiments, the multiple sets of information are collected by the inspection equipment during the welding process. After receiving the multiple sets of image information, the computer processes each set of information into a unified format to obtain the first set of information.

[0051] S22, through normalization, converts the first information into the second information.

[0052] The value of the second piece of information is greater than zero and less than one.

[0053] In some embodiments, the first information is transformed into the range of 0 to 1 using the Max-Min normalization method.

[0054] S23. Based on the second information, the variance coefficient is obtained.

[0055] In this way, the variance coefficients of each group of information can be calculated.

[0056] S24. Based on the variance coefficient, we obtain the capability characterization information.

[0057] Information on specific characterization capabilities is selected based on the calculated variance coefficients.

[0058] S3. Based on the ability representation information, divide the first training set and the first test set.

[0059] S4. Establish a first stress detection model based on the first training set, and optimize the first stress detection model based on the first test set to obtain a second stress detection model.

[0060] Figure 3 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Figure 3 As shown, in some embodiments, there are multiple sets of image information. When performing step S4, the stress measurement method further includes the following steps:

[0061] S41, construct the first convolutional neural network.

[0062] S42, after importing the first training set into the convolutional neural network, the first stress detection model is obtained.

[0063] In some embodiments, a first stress detection model is obtained by learning and training the first training set through input, convolution, pooling, fully connected layers, and output.

[0064] S43, input the first test set into the first stress detection model to obtain the first test result.

[0065] The first test set is input into the first stress detection model, and the first stress detection model is tested using the first test set to obtain the first test result.

[0066] In one implementation, the first test result can be the test accuracy.

[0067] S44. If the first test result is less than the preset expected value, repeat the step of importing the first training set into the convolutional neural network to obtain the first stress detection model.

[0068] Understandably, if the test accuracy meets the expected value, the training of the first stress detection model is stopped, and the second stress detection model is output.

[0069] S5 measures actual stress information.

[0070] In some embodiments, the actual stress information includes the position information of the X80 ring and thermal energy information.

[0071] In some embodiments, historical measurement data can be selected as actual stress information.

[0072] S6, optimize the second stress detection model based on the actual stress information.

[0073] Figure 4 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Figure 4As shown, in some embodiments, there are multiple sets of image information, including the shape of the molten pool and energy distribution. When performing step S6, the stress measurement method further includes the following steps:

[0074] S61, input the actual stress information into the second stress detection model.

[0075] The second stress detection model receives actual stress information. This actual stress information includes the location where the electric arc heat source transfers heat energy to the X80 ring, and the heat energy information itself.

[0076] In some embodiments, the actual stress information includes multiple sets of image information. Any one of these sets is used as the observation data, and the other data besides the observation data are used as the model fitting data.

[0077] S62 describes the first heat flux density distribution using a Gaussian function.

[0078] In some embodiments, the model fitting data is used to calculate the first heat flux density distribution through a Gaussian function.

[0079] In one implementation, the first heat flux density distribution is calculated using a Gaussian function. The formula for calculating the Gaussian function is as follows:

[0080]

[0081] Where q(r) represents the surface heat flow at radius r; q(0) represents the maximum heat flow at the center of the heat source; r represents the distance from the center of the heat source; and c represents the heat source concentration factor.

[0082] In some embodiments, the value of c is related to the welding method.

[0083] In some embodiments, a Gaussian function is used to describe the first heat flux density distribution of the heating spot. The heating spot includes the location where the electric arc heat source transfers heat energy to the X80 ring.

[0084] S63. Based on the image information, a double ellipsoidal heat source model is constructed.

[0085] In one implementation, the calculation formula for the double ellipsoidal heat source model is as follows:

[0086]

[0087] Where q represents heat flux; x, y, and z represent coordinates relative to the center of the heat source; Q1 and Q2 represent the energy densities of the front and rear hemispheres of the heat source; a 1，2 b and c represent parameters related to the shape of the molten pool.

[0088] In some embodiments, the shape of the molten pool and the energy distribution generated on the X80 ring are collected to construct a double ellipsoidal heat source model.

[0089] In one implementation, a double ellipsoidal heat source model is constructed based on data from the model fitting data used to characterize the shape of the molten pool and the energy distribution.

[0090] S64, based on the double ellipsoidal heat source model, yields the second heat flux density distribution.

[0091] The second heat flux distribution is generated by calculating the double ellipsoidal heat source model.

[0092] S65, input the first heat flux density distribution and the second heat flux density distribution into the second stress detection model, and divide the model into a second training set and a second test set.

[0093] In some embodiments, the second stress detection model receives information on the first heat flux density distribution and the second heat flux distribution generated by the Gaussian function and the double ellipsoidal heat source model, and divides the received information into a second training set and a second test set.

[0094] In one implementation, based on a preset learning rate and step size, all possible data samples are listed, and any subset is selected as the second test set, while the remaining subsets are used as the second training set.

[0095] S66, standardize the second training set to obtain training samples.

[0096] S67. Import the training samples into the second stress detection model, and train the second stress detection model through an iterative method to obtain the third stress detection model.

[0097] To improve the model's accuracy, after step S67, in one implementation, the stress measurement method further includes: verifying the accuracy of the third stress detection model based on a second test set. In some embodiments, the evaluation capability of the third stress detection model is calculated by calculating the root mean square error to obtain multiple sets of accuracy parameters. In one example, the combined parameters corresponding to the minimum root mean square error are selected as the optimal parameters within the data interval for replacement.

[0098] In one implementation, the second training set and the second test set defined in step S65 are replaced. The root mean square error (RMSE) calculation is repeated multiple times until a prediction is performed on all data samples, meaning any subset is used as the second training set in the RMSE calculation, and any subset is used as the second test set in the RMSE calculation.

[0099] As described above, by executing steps S1 to S6, the stress measurement method receives the information collected by the detection equipment during the welding process via a computer, preprocesses the information, divides the processed information into training and testing sets, constructs a convolutional neural network, imports the training set into the convolutional neural network, and trains it through input, convolution, pooling, fully connected, and output to obtain a stress detection model. The test set is then used to test the stress detection model until the result meets the expected value. The stress detection model receives the location and heat energy information of the arc heat source transferring heat energy to the X80 ring, and describes the heat flux density distribution of the heating spot using a Gaussian function. It collects the shape and energy distribution of the molten pool generated on the X80 ring to construct a double ellipsoidal heat source model. The stress detection model then receives the heat flux density distribution information calculated by the Gaussian function and the double ellipsoidal heat source model, imports the relevant data into the stress detection model, and plots the residual stress measurement curve. This eliminates the need for manual modeling by staff, facilitating operation and allowing for more intuitive data display, while ensuring the accuracy of residual stress detection.

[0100] To prevent data tampering and ensure data accuracy, Figure 5 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Figure 5 As shown, in some embodiments, after performing step S6, the stress measurement method further includes the following steps:

[0101] S7. Based on the third stress detection model, the measurement results are obtained.

[0102] In some embodiments, the second test set is input into the third stress detection model to obtain the residual stress measurement curve.

[0103] S8 uploads the measurement results to the blockchain storage.

[0104] In some embodiments, when performing step S8, the stress measurement method further includes:

[0105] The measurement results identified in each group are processed into blocks that meet the criteria. Upon joining the network, each node in the blockchain network generates a local public-private key pair as its own identifier in the network. When a node is waiting for its local role to become a candidate node, it broadcasts a leader application to other nodes in the network and sends it.

[0106] One of the nodes verifies the received applications. If the verification passes, it returns the relevant verification data to the candidate nodes and no longer approves new applications. If the verification fails, it requests the entire network consensus to check for errors. If the requesting node is found to be erroneous, it modifies its local records and re-verifies.

[0107] Once a candidate node becomes the leader node, the other nodes become follower nodes. The leader node then broadcasts block record information. After receiving the information, each follower node broadcasts the received information to the other follower nodes and records the number of times it is repeated. After the broadcast ends, each follower node uses the information with the highest number of repetitions to generate a block header and sends a confirmation request to the leader node.

[0108] After receiving the confirmation request, the leader node checks whether the verification data is correct. If any follower node is incorrect, it resends the verification data to that follower node and requests the other follower nodes to send the request information. After all the request information sent by the follower nodes is verified, the leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become a leader node again until the dormant period ends. After receiving the confirmation information, the follower nodes add the newly generated blocks to the chain and return their candidate status.

[0109] The stress measurement method provided in this application process the measurement results of each group of identified blocks into blocks that meet the conditions. When joining the network, each node in the blockchain network generates a local public-private key pair as its own identifier in the network. When a node is waiting for its local role to become a candidate node, it broadcasts a leader application to other nodes in the network and sends it. If the verification is successful, the relevant verification data is returned to the candidate node. When the candidate node becomes the leader node, other nodes become follower nodes. Then the leader node confirms the follower nodes in the block header, and at the same time, the follower node sends a confirmation application to the leader node. The leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become the leader node again until the dormant period ends. After receiving the confirmation information, the follower nodes add the newly generated blocks to the chain and return the candidate identity. This method can upload the measurement data to the blockchain for storage, avoid data tampering, facilitate the accuracy of data verification by subsequent staff, reduce the energy consumption of blockchain data processing, and improve its data processing efficiency.

[0110] When acquiring image information using detection equipment, the equipment's position may not be optimal. This results in inaccurate data, which is detrimental to the generation of stress detection models.

[0111] In order to acquire accurate image information, the stress measurement method provided in this application embodiment can adjust the position of the detection device according to the image information.

[0112] Figure 6 This is a schematic flowchart of one embodiment of the stress measurement method provided in this application. Figure 6 As shown, in some embodiments, after performing step S1 and before performing step S2, the stress measurement method further includes the following steps:

[0113] S101 extracts the image information frame by frame to obtain a toroidal image.

[0114] S102, based on the display ratio of the toroidal image, perform block processing to obtain the image block result.

[0115] S103 extracts the high-frequency components of the image block results through Fourier transform to obtain the high-frequency results of the image.

[0116] For each group of toroidal images after segmentation, Fourier transform is used to analyze and extract high-frequency components from the image segmentation results.

[0117] S104 uses Gaussian filtering to smooth the high-frequency results of the image, resulting in a smoothed image.

[0118] S105 uses an image pyramid to perform scale normalization on the toroidal image in the image smoothing result.

[0119] S106 uses a backbone convolutional neural network to extract features from the toroidal image in the image normalization result, thus obtaining the image feature extraction result.

[0120] S107: Input the image feature extraction results into the bidirectional feature pyramid to obtain the feature fusion results.

[0121] S108 performs classification regression on the feature fusion results, and outputs the detection box and category.

[0122] S109, Generate the coordinates of the detection box.

[0123] The device detection frame information in the toroidal image is collected, and the corresponding detection frame coordinates are generated.

[0124] S110 enlarges and crops the toroidal image to obtain the cropped image.

[0125] S111 filters out simple negative samples in the cropped image to obtain a filtered image set.

[0126] In one implementation, simple negative samples belonging to the background in each group of cropped images are filtered out using RPN.

[0127] S112, classify and regress the regions of the detection device in the filtered image set to obtain the first classification regression result.

[0128] In one implementation, areas that may contain detection equipment are selected for classification and regression.

[0129] S113, generate anchor boxes in the cropped image, and classify and regress the anchor boxes to obtain the second classification regression result.

[0130] In one implementation, a corresponding number of anchor boxes are generated in each cropped image, and these anchor boxes are classified and regressed.

[0131] S114. Based on the first and second classification results, target information is extracted from each group of images through enlarged cropping.

[0132] This allows for analysis of whether the placement of the testing equipment is correct. If incorrect, the placement of the equipment can be adjusted. For example, it can prompt manual adjustment.

[0133] Based on the same concept, this application also provides a stress measurement system. The method corresponding to the stress measurement system can be the stress measurement method in the foregoing embodiments, and its problem-solving principle is similar to that method. The stress measurement system provided in this application includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the methods and / or technical solutions of the various embodiments of this application.

[0134] Another embodiment of this application provides a computer-readable storage medium having computer program instructions stored thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.

[0135] Specifically, this embodiment may employ any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0136] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0137] The program code contained on a computer-readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0138] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0139] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0140] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or page components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0142] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0143] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0144] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0146] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a device claim may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

Claims

1. A stress measurement method, characterized in that, include: Acquire image information, including the shape of the molten pool and energy distribution; The image information is preprocessed to obtain capability characterization information; The capability characterization information is used to characterize the stress experienced by the pipeline structure; Based on the capability representation information, the first training set and the first test set are divided; A first stress detection model is established based on the first training set, and a second stress detection model is obtained by optimizing the first stress detection model based on the first test set. Measure actual stress information, which includes the location where the electric arc heat source transfers heat energy to the pipeline structure and heat energy information; Based on the actual stress information, the second stress detection model is optimized to obtain the third stress detection model; When performing the step of optimizing the second stress detection model based on the actual stress information to obtain the third stress detection model, the stress measurement method further includes: The actual stress information is input into the second stress detection model; The first heat flux density distribution is described by a Gaussian function; Based on the image information, a double ellipsoidal heat source model is constructed; Based on the aforementioned double ellipsoidal heat source model, the second heat flux density distribution is obtained; Input the first heat flux density distribution and the second heat flux density distribution into the second stress detection model, and divide the model into a second training set and a second test set. The second training set is standardized to obtain training samples; The training samples are imported into the second stress detection model, and the second stress detection model is trained through an iterative method to obtain the third stress detection model.

2. The stress measurement method according to claim 1, characterized in that, After performing the steps of measuring the actual stress information and optimizing the second stress detection model based on the actual stress information, the stress measurement method further includes: The measurement results are obtained based on the third stress detection model. The measurement results are uploaded to the blockchain storage.

3. The stress measurement method according to claim 1 or 2, characterized in that, The image information comprises multiple sets. When performing the step of preprocessing the image information to obtain capability characterization information, the stress measurement method further includes: The multiple sets of image information are processed into the same format to obtain the first information; The first information is converted into the second information through normalization; the value of the second information is greater than zero and less than one. Based on the second piece of information, the variance coefficients are obtained; The capability characterization information is obtained based on the variance coefficient.

4. The stress measurement method according to claim 3, characterized in that, When performing the steps of establishing a first stress detection model based on the first training set and optimizing the first stress detection model based on the first test set to obtain a second stress detection model, the stress measurement method further includes: Construct the first convolutional neural network; The first training set is imported into the convolutional neural network to obtain the first stress detection model.

5. The stress measurement method according to claim 4, characterized in that, When performing the steps of establishing a first stress detection model based on the first training set and optimizing the first stress detection model based on the first test set to obtain a second stress detection model, the stress measurement method further includes: Input the first test set into the first stress detection model to obtain the first test result; If the first test result is less than the preset expected value, then repeat the step of importing the first training set into the convolutional neural network to obtain the first stress detection model.

6. The stress measurement method according to claim 2, characterized in that, The image information includes the shape of the molten pool and the energy distribution; after performing the step of optimizing the second stress detection model based on the actual stress information to obtain the third stress detection model, the stress measurement method further includes: The second test set is input into the third stress detection model to obtain the residual stress measurement curve.

7. The stress measurement method according to claim 1 or 2, characterized in that, After performing the step of acquiring image information; Before performing the step of preprocessing the image information to obtain capability characterization information; The stress measurement method further includes: The image information is extracted frame by frame to obtain a toroidal image; Based on the display ratio of the toroidal image, the image is divided into blocks to obtain the image block results; The high-frequency components of the image block results are extracted by Fourier transform to obtain the high-frequency results of the image. Gaussian filtering is used to smooth the high-frequency results of the image, resulting in a smoothed image. The toroidal image in the image smoothing result is scaled and normalized using an image pyramid to obtain the image normalization result. By using a backbone convolutional neural network, features are extracted from the toroidal image in the image normalization result to obtain the image feature extraction result; The image feature extraction results are input into a bidirectional feature pyramid to obtain the feature fusion results; The feature fusion results are then classified and regressed to output detection boxes and categories. Generate the coordinates of the detection box; The toroidal image is enlarged and cropped to obtain a cropped image; Filter the simple negative samples in the cropped images to obtain a set of filtered images; The regions of the detection devices in the filtered image set are classified and regressed to obtain the first classification and regression results; Anchor boxes are generated in the cropped image, and the anchor boxes are classified and regressed to obtain the second classification and regression results; Based on the first and second classification results, target information is extracted from the cropped image through enlarged cropping.

8. A stress measurement system, characterized in that, include: At least one processor; A memory that is communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 7.

9. A computer-readable medium having stored thereon computer program instructions that can be executed by a processor to implement the method as described in any one of claims 1 to 7.

Images