Method for generating prediction model of crush strength of steel pipe, method for predicting crush strength of steel pipe, method for determining manufacturing characteristics of steel pipe, and method for manufacturing steel pipe

Abstract

The present application provides a steel pipe crush strength prediction model generation method, a steel pipe crush strength prediction method, a steel pipe manufacturing characteristic determination method, and a steel pipe manufacturing method, which can accurately predict the crush strength of a steel pipe or a coated steel pipe after forming by considering the pipe forming strain during steel pipe forming. The steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation method is input with the steel pipe manufacturing characteristics of the steel pipe to be predicted, including the steel pipe shape after forming, the steel pipe strength characteristics after forming, and the pipe forming strain during steel pipe forming, to predict the crush strength of the steel pipe after forming (steps S1-S5). In addition, the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation method is input with the steel pipe manufacturing characteristics of the coated steel pipe to be predicted, including the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe forming strain during steel pipe forming, and the coating conditions, to predict the crush strength of the coated steel pipe (steps S11-S15).

Description

Technical Field

[0001] This invention relates to a method for generating a steel pipe crush strength prediction model, a method for predicting the crush strength of steel pipes, a method for determining the manufacturing characteristics of steel pipes, and a method for manufacturing steel pipes. Background Technology

[0002] For steel pipes used in environments subjected to external pressure, there is a possibility of crushing (also known as collapse) due to the external pressure. For example, in submarine pipelines, such crushing can lead to damage to structures and accidents, resulting in significant economic and environmental impacts. Therefore, steel pipes used in applications such as submarine pipelines, which involve high compressive stress, require excellent resistance to crushing.

[0003] Here, as a method for evaluating / predicting the crush resistance performance, for example, there is a known method for predicting / evaluating the crush resistance performance of a target steel pipe as described in Non-Patent Document 1. In Non-Patent Document 1, as a method for predicting / evaluating the crush resistance performance of a target steel pipe, a formula is proposed to predict the crush resistance performance based on the Ovality of the outer diameter shape of the target steel pipe (steel pipe after forming), the yield stress (stress corresponding to 0.5% strain) collected from the center of the material wall thickness or 1 / 4 of the wall thickness (inner surface side), Young's modulus, and Poisson's ratio (D 400 Chapter Local Buckling-External over pressure only 401 Formula (5.10)).

[0004] Non-Patent Literature 1: OFFSHORE STANDARD DNV-OS-F101, SUBMARINE PIPELINE SYSTEMS, DET NORSKE VERITAS, October 2010, SEC5, pp. 41-56

[0005] However, the method for predicting / evaluating the crush resistance of the target steel pipe shown in Non-Patent Document 1 has the following problems.

[0006] In other words, the crushing strength of a steel pipe depends not only on its shape and strength properties (tensile strength, compressive strength, Young's modulus, Poisson's ratio, etc.) after forming, but also on the pipe-forming strain (the strain history during pipe forming). This is because the pipe-forming strain significantly affects the shape and strength properties of the formed steel pipe. On the other hand, in Non-Patent Literature 1, the pipe-forming strain during forming was not considered, resulting in poor accuracy in the predicted crushing strength. The measured crushing strength differed significantly from the predicted value. Therefore, there is a concern that the steel pipe design might be overly conservative, or that crushing due to external pressure lower than predicted could lead to a major accident. Summary of the Invention

[0007] Therefore, this invention was made to solve the above-mentioned existing problems. Its purpose is to provide a method for generating a steel pipe crush strength prediction model, a method for predicting the crush strength of steel pipe, a method for determining the manufacturing characteristics of steel pipe, and a method for manufacturing steel pipe. By considering the pipe forming strain during the forming process, it is possible to predict the crush strength of the formed steel pipe or the coated steel pipe formed by painting after forming with high accuracy.

[0008] To address the aforementioned issues, one aspect of the present invention, a method for generating a steel pipe crush strength prediction model, aims to generate a steel pipe crush strength prediction model that uses machine learning to learn multiple training data to predict the crush strength of a steel pipe after forming. The multiple training data include past steel pipe manufacturing characteristics, such as the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain during forming, as input data. The output data is the past crush strength of the formed steel pipe relative to the input data.

[0009] In addition, the main idea of ​​the steel pipe crush strength prediction method of other aspects of the present invention is to input the steel pipe crush strength prediction model generated by the above-described steel pipe crush strength prediction model generation method with the steel pipe manufacturing characteristics of the steel pipe to be predicted, including the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, so as to predict the crush strength of the steel pipe after forming.

[0010] Furthermore, the main idea of ​​the steel pipe manufacturing characteristic determination method of other aspects of the present invention is to determine the optimal steel pipe manufacturing characteristics by successively changing at least one of the following: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain included in the steel pipe manufacturing characteristics, in order to determine the optimal steel pipe manufacturing characteristics, so that the crushing strength of the steel pipe after forming predicted by the above-mentioned steel pipe crushing strength prediction method is asymptotically close to the target crushing strength of the steel pipe after forming.

[0011] Furthermore, the main idea of ​​another aspect of the steel pipe manufacturing method of the present invention is to include: a steel pipe forming process for forming a steel pipe; a crushing strength prediction process for predicting the crushing strength of the steel pipe formed in the forming process using the above-described steel pipe crushing strength prediction method; and a process for assigning a performance prediction value to the steel pipe formed in the forming process by associating the crushing strength of the steel pipe predicted by the crushing strength prediction process with the performance prediction value.

[0012] Furthermore, the main idea of ​​another aspect of the steel pipe manufacturing method of the present invention is to determine the steel pipe manufacturing conditions based on the optimal steel pipe manufacturing characteristics determined by the above-described steel pipe manufacturing characteristic determination method, and to manufacture the steel pipe under the determined steel pipe manufacturing conditions.

[0013] Furthermore, the main idea of ​​the method for generating a steel pipe crush strength prediction model in other aspects of the present invention is to generate a steel pipe crush strength prediction model that uses machine learning to learn multiple training data to predict the crush strength of a coated steel pipe formed after steel pipe forming and then coated. The multiple training data take past steel pipe manufacturing characteristics, including the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, as input data, and take the past crush strength of the coated steel pipe formed after steel pipe forming relative to the input data as output data.

[0014] Furthermore, the main idea of ​​the other aspects of the present invention for predicting the crush strength of steel pipes is to input the steel pipe crush strength prediction model generated by the above-described method into the steel pipe crush strength prediction model, which includes the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe forming strain during forming, and the coating conditions of the coated steel pipe, in order to predict the crush strength of the coated steel pipe formed after forming.

[0015] Furthermore, the main idea of ​​the steel pipe manufacturing characteristic determination method of other aspects of the present invention is to determine the optimal steel pipe manufacturing characteristics by successively changing at least one of the following: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, in order to make the predicted crush strength of the coated steel pipe asymptotically close to the target crush strength of the coated steel pipe.

[0016] Furthermore, the main idea of ​​another aspect of the steel pipe manufacturing method of the present invention is to include: a coated steel pipe forming step of forming a steel pipe and coating the formed steel pipe to form a coated steel pipe; a crushing strength prediction step of predicting the crushing strength of the coated steel pipe formed in the above-mentioned coated steel pipe forming step using the above-mentioned steel pipe crushing strength prediction method; and a step of assigning a performance prediction value to the coated steel pipe formed in the above-mentioned coated steel pipe forming step by associating the crushing strength of the coated steel pipe predicted by the crushing strength prediction step with the performance prediction value of the coated steel pipe formed in the above-mentioned coated steel pipe forming step.

[0017] Furthermore, the main idea of ​​another method for manufacturing steel pipes according to the present invention is to determine the manufacturing conditions of the coated steel pipe based on the optimal steel pipe manufacturing characteristics determined by the above-described method for determining the manufacturing characteristics of steel pipes, and to manufacture the coated steel pipe under the determined manufacturing conditions of the coated steel pipe.

[0018] The present invention provides a method for generating a steel pipe crush strength prediction model, a method for predicting the crush strength of a steel pipe, a method for determining the manufacturing characteristics of a steel pipe, and a method for manufacturing a steel pipe, which can accurately predict the crush strength of a steel pipe after forming or a coated steel pipe formed by coating after forming by considering the pipe forming strain during forming. Attached Figure Description

[0019] Figure 1 This is a functional block diagram of a simplified structure of a steel pipe manufacturing characteristic determination device, which is a method for generating a steel pipe crush strength prediction model, a method for predicting the crush strength of a steel pipe, and a method for determining the manufacturing characteristics of a steel pipe according to the first and second embodiments of the present invention.

[0020] Figure 2 This is a diagram illustrating the processing flow of the steel pipe crush strength prediction model, which is generated as a neural network model, through the generation method of the steel pipe crush strength prediction model according to the first embodiment of the present invention.

[0021] Figure 3This is a flowchart illustrating the processing flow of the steel pipe manufacturing characteristic calculation unit in the calculation processing unit of the steel pipe manufacturing characteristic determination device applied to the first embodiment of the present invention.

[0022] Figure 4 This is a diagram illustrating the processing flow of a steel pipe crush strength prediction model constructed by a neural network, generated by the method for generating a steel pipe crush strength prediction model according to the second embodiment of the present invention.

[0023] Figure 5 This is a flowchart illustrating the processing flow of the steel pipe manufacturing characteristic calculation unit in the calculation processing unit of the steel pipe manufacturing characteristic determination apparatus, which is used to explain the generation method of the steel pipe crushing strength prediction model, the steel pipe crushing strength prediction method, and the steel pipe manufacturing characteristic determination method according to the second embodiment of the present invention. Detailed Implementation

[0024] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments shown below illustrate apparatus and methods for embodying the technical concept of the present invention. The technical concept of the present invention does not specify the material, shape, structure, arrangement, etc., of the constituent components as described below. Furthermore, the accompanying drawings are schematic. Therefore, it should be noted that the relationship between thickness and planar dimensions, ratios, etc., differ from reality, and even within the drawings, there are parts with different dimensional relationships and ratios.

[0025] (First Implementation)

[0026] Figure 1 A functional block diagram showing a simplified structure of a method for generating a steel pipe crush strength prediction model, a method for predicting the crush strength of a steel pipe, and a device for determining the manufacturing characteristics of a steel pipe, according to the first embodiment of the present invention, is shown.

[0027] Figure 1 The steel pipe manufacturing characteristic determination device 1 shown in the first embodiment generates a steel pipe crush strength prediction model and uses the generated steel pipe crush strength prediction model to predict the crush strength of the formed steel pipe. Furthermore, the steel pipe manufacturing characteristic determination device 1 determines the optimal steel pipe manufacturing characteristics that asymptotically approximate the target crush strength of the formed steel pipe.

[0028] Figure 1The steel pipe manufacturing characteristic determination device 1 shown is a computer system comprising an arithmetic unit 2, an input device 8, a storage device 9, and an output device 10. The arithmetic unit 2, as described below, includes RAM 3, ROM 4, and an arithmetic processing unit 5. These RAM 3, ROM 4, and arithmetic processing unit 5 are connected to the input device 8, storage device 9, and output device 10 via a bus 11. The connection method between the arithmetic unit 2 and the input device 8, storage device 9, and output device 10 is not limited to this; it can be connected wirelessly, or it can be connected using a combination of wired and wireless connections.

[0029] Input device 8 functions as an input port for operators of the system to input various information, such as a keyboard, handwriting pad, touchpad, or mouse. Input devices 8 may contain commands for generating a steel pipe crush strength prediction model, commands for calculating steel pipe manufacturing characteristics, and information on the steel pipe being predicted for crush strength, including the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, the pipe-forming strain during forming, the crush strength of the target formed steel pipe, and the steel pipe manufacturing characteristic determination mode.

[0030] Here, steel pipes are usually manufactured by bending and shaping plate-shaped steel sheets into round tubes, and sometimes they are coated on the surface.

[0031] The steel pipe shape after forming, input into the steel pipe manufacturing characteristics of input device 8, refers to the shape of the steel pipe after the steel plate is formed into a round tube shape. Specifically, the steel pipe shape after forming includes the maximum outer diameter Dmax (mm), the minimum outer diameter Dmin (mm), the average outer diameter Dave (mm), the average plate thickness t (mm), and the roundness (Ovality) fO (%) of the outer diameter shape. The measured values ​​of the steel pipe shape after forming are input into input device 8. This steel pipe shape after forming has a significant impact on the predicted crush strength of the formed steel pipe, therefore, it must be input.

[0032] Furthermore, the crush strength of a steel pipe means the load stress (MPa) at which the steel pipe crushes. Here, "crushing" refers to the state in which the load stress reaches its maximum value and continues to deform until it can no longer maintain its shape relative to external pressure.

[0033] Furthermore, the strength characteristics of the steel pipe after forming signify the strength characteristics of the steel pipe after the steel plate has been formed into a tubular shape. Specifically, the strength characteristics of the steel pipe after forming include the Young's modulus E (GPa), the Poisson's ratio μ (-), the tensile strength YS (MPa), the compressive strength 0.23%YS (stress corresponding to 0.23% strain), and the compressive strength 0.5%YS (stress corresponding to 0.5% strain). These strength characteristics of the steel pipe after forming significantly influence the predicted crushing strength of the formed steel pipe, therefore, must be input. For the strength characteristics of the steel pipe after forming, the input can be either the strength characteristics obtained through finite element analysis simulation based on the strength characteristics of the steel plate before forming, or the measured strength characteristics.

[0034] Furthermore, the forming strain during steel pipe forming is either the tensile strain (%) or compressive strain (%). The forming strain significantly affects the shape and strength characteristics of the formed steel pipe, consequently greatly influencing the predicted crush strength. Therefore, it must be input. The input for the forming strain can be either the strain obtained through forming simulation using finite element analysis based on the strength characteristics of the steel plate before forming, or the measured forming strain.

[0035] In addition, the storage device 9 is, for example, a hard disk drive, a semiconductor drive, an optical drive, etc., and is a device that stores the information required by this system (the information required for the steel pipe crush strength prediction model generation unit 6 and the steel pipe manufacturing characteristic calculation unit 7 to perform their functions, as described later).

[0036] Here, as the information required for the steel pipe crush strength prediction model generation unit 6 to perform its function, for example, multiple training data can be used, which take the past steel pipe manufacturing characteristics, consisting of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, as input data, and the crush strength of the past steel pipe after forming relative to the input data as output data.

[0037] In addition, the information required for the steel pipe manufacturing characteristic calculation unit 7 to perform its functions includes, for example, the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6, the steel pipe manufacturing characteristics of the steel pipe that is input into the steel pipe crush strength prediction model and is input to the input device 8 as the target of the crush strength prediction, which consists of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, the crush strength of the target steel pipe after forming, and the steel pipe manufacturing characteristic determination mode information (information on whether it is the mode that determines the optimal steel pipe manufacturing characteristics).

[0038] Furthermore, the output device 10 functions as an output port for outputting output data from the arithmetic unit 2, such as information on the crush strength (predicted value) of the steel pipe after forming, predicted by the crush strength prediction unit 72 (described later), and information on the optimal steel pipe manufacturing characteristics determined by the steel pipe manufacturing characteristic determination unit 73. The output device 10 may include any display such as a liquid crystal display or an organic display, thereby enabling the display of images based on the output data.

[0039] Next, as Figure 1 As shown, the arithmetic unit 2 includes RAM 3, ROM 4, and a processing unit 5. ROM 4 stores a steel pipe crush strength prediction model generation program 41 and a steel pipe manufacturing characteristic calculation program 42. The processing unit 5 has arithmetic processing functions and is connected to RAM 3 and ROM 4 via bus 11. In addition, RAM 3, ROM 4, and processing unit 5 are connected to input device 8, storage device 9, and output device 10 via bus 11.

[0040] The calculation and processing unit 5 is a functional block that includes a steel pipe crush strength prediction model generation unit 6 and a steel pipe manufacturing characteristic calculation unit 7.

[0041] The steel pipe crush strength prediction model generation unit 6 of the processing unit 5 generates a steel pipe crush strength prediction model using multiple learning data stored in the storage device 9 via machine learning. The multiple learning data take as input data the past steel pipe manufacturing characteristics, consisting of the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain during forming, and output data the crush strength of the formed steel pipe relative to the past crush strength of the formed steel pipe. The machine learning method is a neural network, and the steel pipe crush strength prediction model is a prediction model constructed by the neural network.

[0042] Here, the steel pipe crush strength prediction model generation unit 6, as a functional block, includes a learning data acquisition unit 61, a preprocessing unit 62, a model generation unit 63, and a result storage unit 64. Furthermore, when a steel pipe crush strength prediction model generation command is input to the input device 8 and is received, the steel pipe crush strength prediction model generation unit 6 executes the steel pipe crush strength prediction model generation program 41 stored in the ROM 4, thereby performing the functions of the learning data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64. Each time the steel pipe crush strength prediction model generation unit 6 executes a function, the steel pipe crush strength prediction model is updated.

[0043] The execution of the functions of the learning data acquisition unit 61, preprocessing unit 62, model generation unit 63, and result storage unit 64 of the steel pipe crush strength prediction model generation unit 6 corresponds to the method of generating a steel pipe crush strength prediction model according to the first embodiment of the present invention, which generates a steel pipe crush strength prediction model by machine learning multiple learning data to predict the crush strength of the steel pipe after forming. In this method, the multiple learning data take the past steel pipe manufacturing characteristics, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, as input data, and take the past crush strength of the steel pipe after forming relative to the input data as output data.

[0044] Here, the learning data acquisition unit 61 processes multiple learning data stored in the storage device 9. The multiple learning data take past steel pipe manufacturing characteristics, consisting of the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain during forming, as input data, and the crushing strength of the formed steel pipe relative to the input data as output data. Each learning data consists of a combination of input data and output data.

[0045] Furthermore, the preprocessing unit 62 processes the multiple learning data acquired by the learning data acquisition unit 61 into data for generating a steel pipe crush strength prediction model. Specifically, the preprocessing unit 62 standardizes (normalizes) the data between 0 and 1 so that the neural network model can read past performance information of the steel pipe manufacturing characteristics, which consists of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming.

[0046] Furthermore, the model generation unit 63 performs the following processing: It generates a steel pipe crush strength prediction model from multiple learning data preprocessed by the preprocessing unit 62 using machine learning. The steel pipe crush strength prediction model takes past steel pipe manufacturing characteristics, consisting of the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain during forming, as input data, and outputs the crush strength of the formed steel pipe in the past. In this embodiment, a neural network is used as the machine learning method, and therefore a neural network model is generated as the steel pipe crush strength prediction model. That is, the model generation unit 63 creates a neural network model as the steel pipe crush strength prediction model, which establishes a correlation between the input performance data (past performance data of steel pipe manufacturing characteristics) and the output performance data (past performance data of the crush strength of the formed steel pipe) from the learning data processed for generating the steel pipe crush strength prediction model. The neural network model is represented, for example, by a function.

[0047] Specifically, the model generation unit 63 sets the hyperparameters used in the neural network model and learns the neural network model based on these hyperparameters. Commonly set hyperparameters include the number of hidden layers, the number of neurons in each hidden layer, the dropout rate in each hidden layer, and the activation function in each hidden layer, but these are not limited to these.

[0048] Figure 2 The process flow of the steel pipe crush strength prediction model generated as a neural network model by the method of generating the steel pipe crush strength prediction model according to the first embodiment of the present invention is shown.

[0049] The steel pipe crushing strength prediction model, which is a neural network model, includes an input layer 101, an intermediate layer 102, and an output layer 103 in sequence from the input side.

[0050] When the model generation unit 63 performs learning based on a neural network model using hyperparameters, the input layer 101 stores past steel pipe manufacturing performance information, which consists of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, which constitutes the learning data processed by the preprocessing unit 62. That is, past steel pipe manufacturing performance information standardized between 0 and 1.

[0051] The intermediate layer 102 consists of multiple hidden layers, each containing multiple neurons. The number of hidden layers in the intermediate layer 102 is not particularly limited, but empirically, too many hidden layers will reduce prediction accuracy, so it is preferred to have 5 or fewer layers.

[0052] The output layer 103 combines the information from the neurons transmitted by the intermediate layer 102 and outputs it as the final crush strength of the steel pipe after forming. Based on this output and the read past crush strength data of the steel pipe after forming, the weight coefficients in the neural network model are gradually optimized, thereby learning.

[0053] The result storage unit 64 stores the learning data, the parameters (weight coefficients) of the neural network model, and the output results of the neural network model relative to the learning data in the storage device 9.

[0054] The steel pipe manufacturing characteristic calculation unit 7 of the calculation processing unit 5 performs the following processing: It inputs the steel pipe manufacturing characteristics of the steel pipe to be predicted for crushing strength, which are the steel pipe shape after forming, the strength characteristics of the steel pipe after forming, and the pipe-forming strain during forming, into the steel pipe crushing strength prediction model generated by the steel pipe crushing strength prediction model generation unit 6, to predict the crushing strength of the steel pipe after forming corresponding to these manufacturing characteristics. Furthermore, the steel pipe manufacturing characteristic calculation unit 7 performs the following processing: When the steel pipe manufacturing characteristic determination mode information is a steel pipe manufacturing characteristic determination mode, it successively changes at least one of the steel pipe shape after forming, the strength characteristics of the steel pipe after forming, and the pipe-forming strain during forming, in a manner that the predicted crushing strength of the steel pipe after forming asymptotically approaches the required target crushing strength of the steel pipe after forming, to determine the optimal steel pipe manufacturing characteristics.

[0055] To perform this process, the steel pipe manufacturing characteristic calculation unit 7, as shown in section 7... Figure 1 The unit shown is a functional block comprising an information reading unit 71, a crush strength prediction unit 72, a steel pipe manufacturing characteristic determination unit 73, and a result output unit 74.

[0056] The information reading unit 71 processes the information required for the steel pipe manufacturing characteristic calculation unit 7 to perform its functions, which is stored in the storage device 9. Specifically, the information reading unit 71 processes the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6. In addition, the information reading unit 71 processes the following: it reads the steel pipe manufacturing characteristics of the steel pipe that is the target of the crush strength prediction, which are input into the steel pipe crush strength prediction model, including the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming; the crush strength of the target steel pipe after forming; and the steel pipe manufacturing characteristic determination mode information.

[0057] In addition, the crushing strength prediction unit 72 performs the following processing: it inputs the steel pipe manufacturing characteristics, which are the steel pipe shape after forming, the steel pipe strength characteristics after forming, and the pipe forming strain during forming, which are read by the information reading unit 71 into the steel pipe crushing strength prediction model read by the information reading unit 71, to predict the crushing strength of the steel pipe after forming.

[0058] Furthermore, the steel pipe manufacturing characteristic determination unit 73 and the crushing strength prediction unit 72 perform the following processing: When the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is a steel pipe manufacturing characteristic determination mode, the optimal steel pipe manufacturing characteristics are determined by successively changing at least one of the following: the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe forming strain during forming, in a manner that the predicted crushing strength of the formed steel pipe asymptotically approaches the required target crushing strength of the formed steel pipe. The optimal steel pipe manufacturing characteristics are then output to the result output unit 74. Additionally, the steel pipe manufacturing characteristic determination unit 73 performs the following processing: When the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is not a steel pipe manufacturing characteristic determination mode, the crushing strength information (predicted value) of the formed steel pipe predicted by the crushing strength prediction unit 72 is output to the result output unit 74.

[0059] In addition, the result output unit 74 performs the process of outputting the information of the determined optimal steel pipe manufacturing characteristics or the information of the predicted crush strength of the steel pipe after forming (predicted value) to the output device 10, and performs the process of storing this information in the storage device 9.

[0060] Next, refer to Figure 3 The processing flow of the steel pipe manufacturing characteristic calculation unit 7 in the calculation processing unit 5 of the steel pipe manufacturing characteristic determination device 1 applicable to the first embodiment of the present invention will be described.

[0061] If the steel pipe manufacturing characteristic calculation unit 7 inputs a calculation instruction for steel pipe manufacturing characteristics into the input device 8 and receives the calculation instruction for steel pipe manufacturing characteristics, it executes the steel pipe manufacturing characteristic calculation program 42 stored in the ROM 4, thereby performing the functions of the information reading unit 71, the crushing strength prediction unit 72, the steel pipe manufacturing characteristic determination unit 73, and the result output unit 74.

[0062] First, in step S1, the information reading unit 71 of the steel pipe manufacturing characteristic calculation unit 7 reads the steel pipe crushing strength prediction model generated by the steel pipe crushing strength prediction model generation unit 6, which is stored in the storage device 9.

[0063] Next, in step S2, the information reading unit 71 reads the information on the crushing strength of the steel pipe after forming the desired target steel pipe, which is input from the host (not shown) and stored in the storage device 9.

[0064] Next, in step S3, the information reading unit 71 reads the information of the steel pipe that is the object of the crushing strength prediction, which is input by the operator into the input device 8 and stored in the storage device 9. This information is composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming.

[0065] Next, in step S4, the information reading unit 71 reads the steel pipe manufacturing characteristic determination mode information (information on whether it is the mode that determines the optimal steel pipe manufacturing characteristics) that was input by the operator into the input device 8 and stored in the storage device 9.

[0066] Subsequently, in step S5, the crushing strength prediction unit 72 inputs the steel pipe manufacturing characteristics, which are the steel pipe shape after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, into the steel pipe crushing strength prediction model read in step S1, in order to predict the crushing strength of the steel pipe after forming.

[0067] Steps S1 to S5 correspond to the steel pipe crush strength prediction method of the first embodiment of the present invention. The steel pipe crush strength prediction method inputs the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation method into the steel pipe manufacturing characteristics of the steel pipe after forming, which are the steel pipe shape, the steel pipe strength characteristics after forming, and the pipe forming strain during forming, in order to predict the crush strength of the steel pipe after forming.

[0068] Next, in step S6, the steel pipe manufacturing characteristic determination unit 73 determines whether the steel pipe manufacturing characteristic determination mode information (information on whether it is the mode for determining the optimal steel pipe manufacturing characteristics) read in step S4 is the steel pipe manufacturing characteristic determination mode (the mode for determining the optimal steel pipe manufacturing characteristics).

[0069] Then, if the determination result in step S6 is yes (when it is the steel pipe manufacturing characteristic determination mode), proceed to step S7; if the determination result in step S6 is no (not the steel pipe manufacturing characteristic determination mode), proceed to step S9.

[0070] In step S7, the steel pipe manufacturing characteristic determination unit 73 determines whether the difference between the crush strength (predicted value) of the steel pipe after forming predicted in step S5 and the crush strength (target value) of the desired target steel pipe after forming read in step S2 is within a specified threshold.

[0071] Here, the threshold values ​​specified above vary depending on the target value and manufacturing conditions, but are generally set at 0.5% to 1%.

[0072] Then, if the determination result in step S7 is yes (when it is determined that the difference between the predicted value and the target value is within the specified threshold), proceed to step S8; if the determination result in step S7 is no (when it is determined that the difference between the predicted value and the target value is greater than the specified threshold), proceed to step S10.

[0073] In step S10, the steel pipe manufacturing characteristic determination unit 73 changes at least one of the steel pipe manufacturing characteristics of the steel pipe that is the target of the crushing strength prediction, such as the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, which were read in step S3, and returns to step S5.

[0074] If the process returns to step S5, the crushing strength prediction unit 72 inputs the steel pipe manufacturing characteristics of the steel pipe whose shape, strength characteristics, and pipe-forming strain were changed in step S10 to the steel pipe crushing strength prediction model read in step S1, to predict the crushing strength of the steel pipe after forming again. Then, after step S6, the steel pipe manufacturing characteristic determination unit 73 determines in step S7 whether the difference between the crushing strength (predicted value) of the steel pipe after forming predicted again in step S5 and the target crushing strength (target value) of the steel pipe after forming read in step S2 is within a specified threshold. Then, the series of steps S10, S5, S6, and S7 are repeatedly executed until the determination result is yes.

[0075] On the other hand, if the determination result in step S7 is yes (when it is determined that the difference between the predicted value and the target value is within a specified threshold), the process moves to step S8. In step S8, the steel pipe manufacturing characteristic determination unit 73 determines the optimal steel pipe manufacturing characteristics as the steel pipe manufacturing characteristics determined when the difference between the predicted value and the target value is within a specified threshold, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming.

[0076] Steps S6, S7, S10, S5, S6, S7, and S8 correspond to the steel pipe manufacturing characteristic determination method of the first embodiment of the present invention. The steel pipe manufacturing characteristic determination method determines the optimal steel pipe manufacturing characteristics by successively changing at least one of the following factors included in the steel pipe manufacturing characteristics: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, in order to determine the optimal steel pipe manufacturing characteristics.

[0077] Then, in step S9, if the result output unit 74 of the steel pipe manufacturing characteristic calculation unit 7 determines the optimal steel pipe manufacturing characteristics determined in step S8 (when it is the steel pipe manufacturing characteristic determination mode), it outputs the information of the optimal steel pipe manufacturing characteristics determined in step S8 to the output device 10. On the other hand, if the result output unit 74 determines the negative result in step S6 (when it is not the steel pipe manufacturing characteristic determination mode), it outputs the information of the crushing strength of the steel pipe after forming (predicted value) predicted in step S5 to the output device 10.

[0078] Thus, the processing of the steel pipe manufacturing characteristic calculation unit 7 is completed.

[0079] In this way, the method for generating a steel pipe crush strength prediction model according to the first embodiment of the present invention generates a steel pipe crush strength prediction model (steel pipe crush strength prediction model generation unit 6) that uses multiple learning data to predict the crush strength of the steel pipe after forming by machine learning. The multiple learning data take the past steel pipe manufacturing characteristics, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, as input data, and take the crush strength of the past steel pipe after forming relative to the input data as output data.

[0080] Therefore, it is possible to appropriately generate a steel pipe crush strength prediction model that accurately predicts the crush strength of the steel pipe after forming by taking into account the pipe forming strain during the forming process.

[0081] In addition, the steel pipe crush strength prediction method of the first embodiment of the present invention inputs the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation method into the steel pipe manufacturing characteristics of the steel pipe to be predicted, which are the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain during forming, to predict the crush strength of the steel pipe after forming (steps S1 to S5).

[0082] Therefore, the crushing strength of the steel pipe after forming can be predicted with high accuracy by taking into account the pipe forming strain during the forming process.

[0083] Furthermore, the steel pipe manufacturing characteristic determination method of the first embodiment of the present invention determines the optimal steel pipe manufacturing characteristics by successively changing at least one of the following factors included in the steel pipe manufacturing characteristics: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe-forming strain during forming, in order to determine the optimal steel pipe manufacturing characteristics (steps S6, S7, S10, S5, S6, S7, and S8).

[0084] Therefore, the optimal steel pipe manufacturing characteristics, which are composed of the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain during the forming process, can determine when the predicted crush strength of the formed steel pipe asymptotically approaches the target crush strength of the formed steel pipe.

[0085] In addition, when manufacturing steel pipes, the information (predicted value) of the crush strength of the steel pipe after forming, which is predicted in step S5 and output by the output device 10, can be associated with the steel pipe formed in the forming process.

[0086] In other words, the steel pipe manufacturing method of the first embodiment of the present invention may also include: a steel pipe forming process for forming a steel pipe; a crushing strength prediction process for predicting the crushing strength of the steel pipe formed in the forming process by a steel pipe crushing strength prediction method (steps S1 to S5); and a process for assigning a performance prediction value to the steel pipe formed in the forming process by associating the crushing strength of the steel pipe predicted by the crushing strength prediction process with the performance prediction value of the steel pipe formed in the forming process.

[0087] Here, the correlation between the predicted crush strength of the steel pipe in the process and the formed steel pipe is achieved, for example, by assigning the predicted crush strength (predicted value) of the steel pipe to the formed steel pipe using a mark, or by attaching a label containing the predicted crush strength (predicted value) of the steel pipe to the formed steel pipe.

[0088] Therefore, the person handling the formed steel pipe can determine the crushing strength (predicted value) of the steel pipe.

[0089] In addition, when manufacturing steel pipes, the manufacturing conditions of steel pipes (selection of pipe-making method, bending rate during pipe-making, strain applied during pipe-making, etc.) can be determined based on the information of the optimal steel pipe manufacturing characteristics determined in step S8 output by the output device 10, and the steel pipes can be manufactured under the determined manufacturing conditions.

[0090] In other words, the steel pipe manufacturing method of the first embodiment of the present invention can also determine the steel pipe manufacturing conditions based on the optimal steel pipe manufacturing characteristics determined by the steel pipe manufacturing characteristic determination method (steps S6, S7, S10, S5, S6, S7 and S8), and manufacture the steel pipe under the determined steel pipe manufacturing conditions.

[0091] As a result, the manufactured steel pipe meets the determined optimal steel pipe manufacturing characteristics. Consequently, the predicted crush strength (predicted value) of the steel pipe asymptotically approaches the target crush strength of the formed steel pipe, making it a steel pipe with excellent crush resistance, which can avoid damage and destruction accidents to structures.

[0092] (Second Implementation)

[0093] Reference Figure 1 , Figure 4 and Figure 5 The method for generating the steel pipe crush strength prediction model, the method for predicting the crush strength of the steel pipe, the method for determining the manufacturing characteristics of the steel pipe, and the method for manufacturing the steel pipe according to the second embodiment of the present invention will be described. There are instances where descriptions of components already described in the first embodiment are omitted.

[0094] Figure 1 The steel pipe manufacturing characteristic determination device 1 shown is also applicable to the steel pipe crush strength prediction model generation method, steel pipe crush strength prediction method, and steel pipe manufacturing characteristic determination method of the second embodiment. The steel pipe crush strength prediction model generation method of the second embodiment generates a steel pipe crush strength prediction model for a coated steel pipe formed after forming. The steel pipe crush strength prediction method of the second embodiment uses the generated steel pipe crush strength prediction model to predict the crush strength of the coated steel pipe formed after forming. The steel pipe manufacturing characteristic determination method of the second embodiment determines the optimal steel pipe manufacturing characteristics such that the predicted crush strength of the coated steel pipe asymptotically approaches the desired target crush strength of the coated steel pipe.

[0095] Figure 1 The steel pipe manufacturing characteristic determination device 1 of the second embodiment shown is a computer system equipped with a computing device 2, an input device 8, a storage device 9 and an output device 10. Its basic structure has been described, so the description is omitted appropriately.

[0096] Similar to the first embodiment, the input device 8 receives commands for generating a steel pipe crush strength prediction model and for calculating steel pipe manufacturing characteristics. In the second embodiment, unlike the first embodiment, the crush strength of a coated steel pipe formed after pipe forming is predicted. Therefore, as steel pipe manufacturing characteristics, in addition to the shape of the coated steel pipe after forming, the strength characteristics of the formed steel pipe, and the pipe-forming strain during forming, coating conditions are also input. Furthermore, the input device 8 receives commands for the crush strength of a coated steel pipe formed after the target steel pipe is formed.

[0097] Here, the shape of the formed steel pipe, its strength characteristics, and the pipe-forming strain are the same as in the first embodiment, but the coating conditions are the highest temperature (°C) and holding time (min) during coating. These coating conditions are input from the measured coating conditions.

[0098] The coating of formed steel pipes is for corrosion protection, especially to ensure excellent corrosion resistance for steel pipes used in submarine pipelines. Coating is typically performed after forming. The coating conditions (maximum temperature (°C) and holding time (min)) affect the strength characteristics of the formed steel pipe, thus directly impacting its crush resistance. Therefore, this information must be input into input device 8. The material properties of the steel pipe change due to the heating of the coating (dislocation accumulation / recovery / strain aging, etc.), thereby increasing or decreasing its crush strength based on the crush strength of the formed steel pipe (crush resistance before coating).

[0099] In addition, the storage device 9 is a device that stores the information required for the steel pipe crush strength prediction model generation unit 6 and the steel pipe manufacturing characteristic calculation unit 7 to perform their functions. As the information required for the steel pipe crush strength prediction model generation unit 6 to perform its functions, it can include multiple learning data that take the past steel pipe manufacturing characteristics, which consist of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, as input data, and take the crush strength of the coated steel pipe formed after forming and coated in the past relative to the input data as output data.

[0100] In addition, the information required for the steel pipe manufacturing characteristic calculation unit 7 to perform its functions includes: the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6; the steel pipe manufacturing characteristics input to the steel pipe crush strength prediction model and input to the input device 8, which are the objects of the crush strength prediction of the coated steel pipe; the steel pipe shape after forming; the steel pipe strength characteristics after forming; the pipe forming strain during forming; and the coating conditions; the crush strength of the coated steel pipe formed after forming the target steel pipe and then coating it; and the steel pipe manufacturing characteristic determination mode information (information on whether it is the mode that determines the optimal steel pipe manufacturing characteristics).

[0101] In addition, the output device 10 functions as an output port for outputting output data from the computing device 2, such as information on the crush strength (predicted value) of a coated steel pipe formed by coating after steel pipe forming, predicted by the crush strength prediction unit 72, and information on the optimal steel pipe manufacturing characteristics determined by the steel pipe manufacturing characteristics determination unit 73.

[0102] Next, the computing device 2 has the same structure as the first embodiment, and in particular, it includes a steel pipe crush strength prediction model generation unit 6 and a steel pipe manufacturing characteristic computing unit 7 as functional blocks.

[0103] The steel pipe crush strength prediction model generation unit 6 of the processing unit 5 generates a steel pipe crush strength prediction model using multiple learning data stored in the storage device 9 via machine learning. The multiple learning data take past steel pipe manufacturing characteristics, consisting of the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, the pipe-forming strain during forming, and the coating conditions, as input data. The output data is the crush strength of a coated steel pipe formed after forming, relative to the past input data. The machine learning method is the same as in the first embodiment, which is a neural network. The steel pipe crush strength prediction model is a prediction model constructed by the neural network.

[0104] Here, the steel pipe crush strength prediction model generation unit 6, similar to that in the first embodiment, includes a learning data acquisition unit 61, a preprocessing unit 62, a model generation unit 63, and a result storage unit 64 as functional blocks. Furthermore, when a steel pipe crush strength prediction model generation command is input to the input device 8 and is received, the steel pipe crush strength prediction model generation unit 6 executes the steel pipe crush strength prediction model generation program 41 stored in the ROM 4, thereby performing the functions of the learning data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64. Each time the steel pipe crush strength prediction model generation unit 6 performs a function, the steel pipe crush strength prediction model is updated.

[0105] The execution processing of each function of the learning data acquisition unit 61, preprocessing unit 62, model generation unit 63, and result storage unit 64 based on the steel pipe crush strength prediction model generation unit 6 corresponds to the method of generating a steel pipe crush strength prediction model in the second embodiment of the present invention, which uses machine learning to generate multiple learning data to predict the crush strength of a coated steel pipe formed after steel pipe forming and then coated. In this method, the multiple learning data take the past steel pipe manufacturing characteristics, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, as input data, and take the past crush strength of the coated steel pipe formed after forming and then coated relative to the input data as output data.

[0106] Here, the learning data acquisition unit 61 processes multiple learning data stored in the storage device 9. The multiple learning data take as input data the past steel pipe manufacturing characteristics, which consist of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe-forming strain during forming, and the coating conditions. The output data is the crushing strength of a coated steel pipe formed after forming and coated, relative to the past input data. Each learning data consists of a combination of input data and output data.

[0107] In addition, similarly to the first embodiment, the preprocessing unit 62 processes the multiple learning data acquired by the learning data acquisition unit 61 into data for generating a steel pipe crushing strength prediction model.

[0108] Furthermore, the model generation unit 63 performs the following processing: It generates a steel pipe crush strength prediction model using multiple learning data preprocessed by the preprocessing unit 62 through machine learning. The steel pipe crush strength prediction model takes as input data the past steel pipe manufacturing characteristics, consisting of the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe-forming strain during forming, and the coating conditions. It outputs as the crush strength of past coated steel pipes formed after forming. In this embodiment, a neural network is used as the machine learning method, similar to the first embodiment, and thus a neural network model is generated as the steel pipe crush strength prediction model.

[0109] Figure 4 The process flow of the steel pipe crush strength prediction model generated as a neural network model by the method of generating the steel pipe crush strength prediction model according to the second embodiment of the present invention is shown.

[0110] The steel pipe crushing strength prediction model, which is a neural network model, includes an input layer 101, an intermediate layer 102, and an output layer 103 in sequence from the input side.

[0111] When the model generation unit 63 performs learning based on a neural network model using hyperparameters, the input layer 101 stores past steel pipe manufacturing performance information, which consists of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, which constitute the learning data processed by the preprocessing unit 62. That is, past steel pipe manufacturing performance information standardized between 0 and 1.

[0112] The intermediate layer 102 consists of multiple hidden layers, each containing multiple neurons.

[0113] The output layer 103 combines the information from the neurons transmitted by the intermediate layer 102 and outputs it as the crush strength of the coated steel pipe after final forming and coating. Based on this output and the read past crush strength data of coated steel pipes, the weight coefficients in the neural network model are gradually optimized, thereby learning.

[0114] The result storage unit 64 stores the learning data, the parameters (weight coefficients) of the neural network model, and the output results of the neural network model relative to the learning data in the storage device 9.

[0115] The steel pipe manufacturing characteristic calculation unit 7 of the calculation processing unit 5 performs the following processing: It inputs the steel pipe manufacturing characteristics of the coated steel pipe, which is the object of the crushing strength prediction, into the steel pipe crushing strength prediction model generated by the steel pipe crushing strength prediction model generation unit 6. These characteristics consist of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe-forming strain during forming, and the coating conditions. The calculation unit predicts the crushing strength of the coated steel pipe formed after forming, which corresponds to these steel pipe manufacturing characteristics. Furthermore, the steel pipe manufacturing characteristic calculation unit 7 performs the following processing: When the steel pipe manufacturing characteristic determination mode information is a steel pipe manufacturing characteristic determination mode, it successively changes at least one of the following factors—the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe-forming strain during forming, and the coating conditions—to determine the optimal steel pipe manufacturing characteristics, in a manner that the predicted crushing strength of the coated steel pipe asymptotically approaches the target crushing strength of the coated steel pipe.

[0116] To perform this process, the steel pipe manufacturing characteristic calculation unit 7, as shown in section 7... Figure 1 The unit shown is a functional block comprising an information reading unit 71, a crush strength prediction unit 72, a steel pipe manufacturing characteristic determination unit 73, and a result output unit 74.

[0117] The information reading unit 71 processes the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6. Additionally, the information reading unit 71 processes the following: it reads information about the steel pipe that is the target of the crush strength prediction, which is input into the steel pipe crush strength prediction model. This information includes the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, as well as information about the crush strength of the target coated steel pipe and the steel pipe manufacturing characteristic determination mode.

[0118] In addition, the crushing strength prediction unit 72 performs the following processing: it inputs the steel pipe manufacturing characteristics, which are the objects of the crushing strength prediction and are read by the information reading unit 71, consisting of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, into the steel pipe crushing strength prediction model read by the information reading unit 71, so as to predict the crushing strength of the coated steel pipe that is coated after forming.

[0119] Furthermore, the steel pipe manufacturing characteristic determination unit 73 and the crushing strength prediction unit 72 perform the following processing: When the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is a steel pipe manufacturing characteristic determination mode, the optimal steel pipe manufacturing characteristics are determined by successively changing at least one of the following: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, in a manner that the predicted crushing strength of the coated steel pipe asymptotically approaches the required target crushing strength of the coated steel pipe. The optimal steel pipe manufacturing characteristics are then output to the result output unit 74. Additionally, the steel pipe manufacturing characteristic determination unit 73 performs the following processing: When the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is not a steel pipe manufacturing characteristic determination mode, the crushing strength information (predicted value) of the coated steel pipe predicted by the crushing strength prediction unit 72 is output to the result output unit 74.

[0120] In addition, the result output unit 74 performs the process of outputting the information of the determined optimal steel pipe manufacturing characteristics or the information of the predicted crush strength of the coated steel pipe (predicted value) to the output device 10, and performs the process of storing this information in the storage device 9.

[0121] Next, refer to Figure 5 The processing flow of the steel pipe manufacturing characteristic calculation unit 7 in the calculation processing unit 5 of the steel pipe manufacturing characteristic determination device 1 applicable to the second embodiment of the present invention will be described.

[0122] If the steel pipe manufacturing characteristic calculation unit 7 inputs a calculation instruction for steel pipe manufacturing characteristics into the input device 8 and receives the calculation instruction for steel pipe manufacturing characteristics, it executes the steel pipe manufacturing characteristic calculation program 42 stored in the ROM 4, thereby performing the functions of the information reading unit 71, the crushing strength prediction unit 72, the steel pipe manufacturing characteristic determination unit 73, and the result output unit 74.

[0123] First, in step S11, the information reading unit 71 of the steel pipe manufacturing characteristic calculation unit 7 reads the steel pipe crushing strength prediction model generated by the steel pipe crushing strength prediction model generation unit 6, which is stored in the storage device 9.

[0124] Next, in step S12, the information reading unit 71 reads the required crush strength information of the coated steel pipe formed by coating after the target steel pipe is formed, which is input by the host (not shown) and stored in the storage device 9.

[0125] Next, in step S13, the information reading unit 71 reads the information input by the operator to the input device 8 and stored in the storage device 9, which is the information of the coated steel pipe that is the object of the crushing strength prediction model. This information is composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions.

[0126] Next, in step S14, the information reading unit 71 reads the steel pipe manufacturing characteristic determination mode information (information on whether it is the mode that determines the optimal steel pipe manufacturing characteristics) that was input by the operator into the input device 8 and stored in the storage device 9.

[0127] Subsequently, in step S15, the crushing strength prediction unit 72 inputs the steel pipe manufacturing characteristics, which are the object of the crushing strength prediction read in step S13, into the steel pipe crushing strength prediction model read in step S11. These characteristics consist of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, in order to predict the crushing strength of the coated steel pipe.

[0128] Steps S11 to S15 correspond to the steel pipe crush strength prediction method of the second embodiment of the present invention. The steel pipe crush strength prediction method inputs the steel pipe manufacturing characteristics, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, into the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation method, in order to predict the crush strength of the coated steel pipe.

[0129] Next, in step S16, the steel pipe manufacturing characteristic determination unit 73 determines whether the steel pipe manufacturing characteristic determination mode information (information on whether it is the mode for determining the optimal steel pipe manufacturing characteristics) read in step S14 is a steel pipe manufacturing characteristic determination mode (the mode for determining the optimal steel pipe manufacturing characteristics).

[0130] Then, if the determination result in step S16 is yes (when it is the steel pipe manufacturing characteristic determination mode), proceed to step S17; if the determination result in step S16 is no (when it is not the steel pipe manufacturing characteristic determination mode), proceed to step S19.

[0131] In step S17, the steel pipe manufacturing characteristic determination unit 73 determines whether the difference between the crushing strength (predicted value) of the coated steel pipe predicted in step S15 and the crushing strength (target value) of the desired target coated steel pipe read in step S12 is within a specified threshold.

[0132] Here, the threshold for this regulation is roughly set at 0.5% to 1%.

[0133] Then, if the determination result in step S17 is yes (when it is determined that the difference between the predicted value and the target value is within the specified threshold), proceed to step S18; if the determination result in step S17 is no (when it is determined that the difference between the predicted value and the target value is greater than the specified threshold), proceed to step S20.

[0134] In step S20, the steel pipe manufacturing characteristic determination unit 73 changes at least one of the following steel pipe manufacturing characteristics read in step S13: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, and returns to step S15.

[0135] If the process returns to step S15, the crushing strength prediction unit 72 inputs the steel pipe manufacturing characteristics of the steel pipe whose shape, strength characteristics, forming strain, and coating conditions were changed in step S20 into the steel pipe crushing strength prediction model read in step S11, to predict the crushing strength of the coated steel pipe again. Then, after step S16, the steel pipe manufacturing characteristic determination unit 73 determines in step S17 whether the difference between the crushing strength (predicted value) of the coated steel pipe predicted again in step S15 and the crushing strength (target value) of the desired target coated steel pipe read in step S12 is within a specified threshold. Then, steps S20, S15, S16, and S17 are repeatedly executed until the determination result is yes.

[0136] On the other hand, if the determination result in step S17 is yes (when it is determined that the difference between the predicted value and the target value is within a specified threshold), the process moves to step S18. In step S18, the steel pipe manufacturing characteristic determination unit 73 determines the optimal steel pipe manufacturing characteristics as the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe forming strain during forming, and the coating conditions when it is determined that the difference between the predicted value and the target value is within a specified threshold.

[0137] Steps S16, S17, S20, S15, S16, S17, and S18 correspond to the steel pipe manufacturing characteristic determination method of the second embodiment of the present invention. This steel pipe manufacturing characteristic determination method determines the optimal steel pipe manufacturing characteristics by successively changing at least one of the following factors included in the steel pipe manufacturing characteristics: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions.

[0138] Then, in step S19, when the result output unit 74 of the steel pipe manufacturing characteristic calculation unit 7 determines the optimal steel pipe manufacturing characteristics determined in step S18 (when it is the steel pipe manufacturing characteristic determination mode), it outputs the information to the output device 10. On the other hand, when the result output unit 74 determines the optimal steel pipe manufacturing characteristics determined in step S16 (when it is not the steel pipe manufacturing characteristic determination mode), it outputs the information (predicted value) of the crushing strength of the coated steel pipe formed by coating after steel pipe forming, which was predicted in step S15, to the output device 10.

[0139] Thus, the processing of the steel pipe manufacturing characteristic calculation unit 7 is completed.

[0140] In this way, the method for generating a steel pipe crush strength prediction model according to the second embodiment of the present invention generates a steel pipe crush strength prediction model (steel pipe crush strength prediction model generation unit 6) that uses multiple learning data to predict the crush strength of the steel pipe after forming by machine learning. The multiple learning data take the past steel pipe manufacturing characteristics, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, as input data, and take the crush strength of the past steel pipe after forming relative to the input data as output data.

[0141] Therefore, it is possible to appropriately generate a steel pipe crush strength prediction model that can accurately predict the crush strength of coated steel pipes after coating by taking into account the pipe forming strain during steel pipe forming.

[0142] In addition, when generating the steel pipe crush strength prediction model to predict the crush strength of coated steel pipes, the coating conditions that have a significant impact on the crush strength of coated steel pipes are also considered, which can make the steel pipe crush strength prediction model more accurate.

[0143] In addition, the steel pipe crush strength prediction method of the second embodiment of the present invention inputs the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation method into the steel pipe manufacturing characteristics of the coated steel pipe that is the prediction object, which are composed of the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, to predict the crush strength of the steel pipe after forming (steps S11 to S15).

[0144] Therefore, by taking into account the pipe forming strain during the forming process, the crushing strength of the coated steel pipe, which is then coated after forming, can be predicted with high accuracy.

[0145] Furthermore, when predicting the crush strength of coated steel pipes, the coating conditions that have a significant impact on the crush strength of coated steel pipes are also considered, thus further improving the prediction accuracy of the crush strength of coated steel pipes.

[0146] Furthermore, the steel pipe manufacturing characteristic determination method of the second embodiment of the present invention determines the optimal steel pipe manufacturing characteristics by successively changing at least one of the following factors included in the steel pipe manufacturing characteristics: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions (steps S16, S17, S20, S15, S16, S17, and S18) in a manner that the predicted crush strength of the coated steel pipe asymptotically approaches the target crush strength of the coated steel pipe.

[0147] Therefore, the optimal steel pipe manufacturing characteristics, which are composed of the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe forming strain during forming, and the coating conditions, can determine the predicted crush strength of the coated steel pipe asymptotically close to the target crush strength.

[0148] In addition, when manufacturing coated steel pipes, the information (predicted value) of the crushing strength of the coated steel pipe predicted in step S15, output by the output device 10, can be associated with the coated steel pipe formed in the forming process.

[0149] In other words, the steel pipe manufacturing method of the second embodiment of the present invention may also include: a coated steel pipe forming step of forming a steel pipe and coating the formed steel pipe to form a coated steel pipe; a crushing strength prediction step of predicting the crushing strength of the coated steel pipe formed in the coated steel pipe forming step by means of a steel pipe crushing strength prediction method (steps S11 to S15); and a step of assigning a performance prediction value to the coated steel pipe formed in the coated steel pipe forming step by associating the crushing strength of the coated steel pipe predicted by the crushing strength prediction step with the performance prediction value of the coated steel pipe formed in the coated steel pipe forming step.

[0150] Here, the correlation between the predicted crush strength of the coated steel pipe and the predicted crush strength of the coated steel pipe in the process is achieved, for example, by using a mark to assign the predicted crush strength (predicted value) of the coated steel pipe to the coated steel pipe, or by affixing a label to the coated steel pipe that records the predicted crush strength (predicted value) of the coated steel pipe.

[0151] Therefore, those who handle coated steel pipes can determine the crushing strength (predicted value) of the coated steel pipe.

[0152] In addition, when manufacturing coated steel pipes, the manufacturing conditions of coated steel pipes (selection of pipe-making method, bending rate during pipe-making, strain during pipe-making, heating rate during coating, maximum temperature reached during coating, holding time of maximum temperature reached during coating, cooling rate after holding time of maximum temperature reached during coating, etc.) can be determined based on the information of the optimal steel pipe manufacturing characteristics determined in step S18 output by the output device 10. Coated steel pipes are manufactured under the determined manufacturing conditions of coated steel pipes.

[0153] In other words, the steel pipe manufacturing method of the second embodiment of the present invention can also determine the manufacturing conditions of the coated steel pipe based on the optimal steel pipe manufacturing characteristics determined by the method for determining the manufacturing characteristics of coated steel pipe (steps S16, S17, S20, S15, S16, S17 and S18), and manufacture the coated steel pipe under the determined manufacturing conditions of the coated steel pipe.

[0154] Therefore, the manufactured coated steel pipe meets the determined optimal steel pipe manufacturing characteristics. As a result, the predicted crush strength (predicted value) of the coated steel pipe asymptotically approaches the required target crush strength of the coated steel pipe, becoming a coated steel pipe with excellent crush resistance, which can avoid damage and destruction accidents to structures.

[0155] The embodiments of the present invention have been described above, but the present invention is not limited thereto and various changes and improvements can be made.

[0156] For example, in the method for generating a steel pipe crush strength prediction model according to the first embodiment, the past steel pipe manufacturing characteristics that become input data when generating the steel pipe crush strength prediction model include the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe-forming strain during forming. However, the past steel pipe manufacturing characteristics only need to include the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe-forming strain during forming, or they may include other past steel pipe manufacturing characteristics, such as the strength characteristics of the steel plate before forming.

[0157] Even in the method for generating the steel pipe crush strength prediction model in the second embodiment, when generating the steel pipe crush strength prediction model, the past steel pipe manufacturing characteristics that become the input data only need to include the shape of the steel pipe after it has been formed, the strength characteristics of the steel pipe after it has been formed, the pipe forming strain and coating conditions during the forming of the steel pipe, or other past steel pipe manufacturing characteristics, such as the strength characteristics of the steel plate before the steel pipe was formed.

[0158] Furthermore, in the method for generating the steel pipe crush strength prediction model in the first and second embodiments, when generating the steel pipe crush strength prediction model, the shape of the steel pipe after past forming, which becomes the input data, is not limited to the maximum outer diameter Dmax (mm), the minimum outer diameter Dmin (mm), the average outer diameter Dave (mm), the average plate thickness t (mm), and the roundness (Ovality) fO (%) of the outer diameter shape of the steel pipe.

[0159] Furthermore, in the method for generating the steel pipe crush strength prediction model in the first and second embodiments, when generating the steel pipe crush strength prediction model, the strength characteristics of the steel pipe after forming, which are used as input data, are not limited to the Young's modulus E (GPa), the Poisson's ratio μ (-), the tensile strength YS (MPa), the compressive strength 0.23%YS (stress corresponding to 0.23% strain), and the compressive strength 0.5%YS (stress corresponding to 0.5% strain).

[0160] Furthermore, in the steel pipe crush strength prediction method of the first embodiment, the steel pipe crush strength prediction model is input with the steel pipe manufacturing characteristics, which consist of the steel pipe shape after forming, the steel pipe strength characteristics after forming, and the pipe-forming strain during forming. These steel pipe manufacturing characteristics only need to include the steel pipe shape after forming, the steel pipe strength characteristics after forming, and the pipe-forming strain during forming; alternatively, other steel pipe manufacturing characteristics, such as the strength characteristics of the steel plate before forming, can also be input.

[0161] Furthermore, even in the steel pipe crush strength prediction method of the second embodiment, the steel pipe crush strength prediction model is input with the steel pipe manufacturing characteristics of the coated steel pipe to be predicted, which consist of the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe-forming strain during forming, and the coating conditions. These steel pipe manufacturing characteristics only need to include the steel pipe shape after forming, the steel pipe strength characteristics after forming, the pipe-forming strain during forming, and the coating conditions; alternatively, other steel pipe manufacturing characteristics, such as the strength characteristics of the steel plate before forming, can also be input.

[0162] Furthermore, in the steel pipe crush strength prediction method of the first and second embodiments, the shape of the steel pipe after forming, which is input into the steel pipe crush strength prediction model, is not limited to the maximum outer diameter Dmax (mm), the minimum outer diameter Dmin (mm), the average outer diameter Dave (mm), the average plate thickness t (mm), and the roundness (Ovality) fO (%) of the outer diameter shape of the steel pipe.

[0163] Furthermore, in the steel pipe crush strength prediction methods of the first and second embodiments, the steel pipe strength characteristics after forming that are input into the steel pipe crush strength prediction model are not limited to the Young's modulus E (GPa), Poisson's ratio μ (-), tensile strength YS (MPa), compressive strength 0.23%YS (stress corresponding to 0.23% strain), and compressive strength 0.5%YS (stress corresponding to 0.5% strain).

[0164] In addition, in the first and second embodiments, the machine learning method is a neural network. The steel pipe crush strength prediction model is a prediction model constructed by a neural network. However, any machine learning method is acceptable, such as a decision tree.

[0165] Example

[0166] To verify the effectiveness of the present invention, the crushing strength of the steel pipe was predicted under the conditions shown in Table 1.

[0167] [Table 1]

[0168]

[0169] In Examples 1 and 2, a steel pipe crush strength prediction model is generated using machine learning with multiple training data. The multiple training data take the past steel pipe shape after forming (maximum outer diameter Dmax (mm), minimum outer diameter Dmin (mm), average outer diameter Dave (mm), average plate thickness t (mm), and roundness (Ovality) fO (%)), the past steel pipe strength characteristics after forming (Young's modulus E (GPa), Poisson's ratio μ (-), tensile strength YS (MPa), compressive strength 0.23%YS (stress corresponding to 0.23% strain) and compressive strength 0.5%YS (stress corresponding to 0.5% strain)) and the pipe forming strain during the past forming process (tensile strain (%)) as input data, and the past crush strength (MPa) of the formed steel pipe relative to the input data as output data.

[0170] Furthermore, in Examples 1 and 2, the steel pipe shape after forming (maximum outer diameter Dmax (mm), minimum outer diameter Dmin (mm), average outer diameter Dave (mm), average plate thickness t (mm), roundness (Ovality) fO (%)), strength characteristics of the formed steel pipe (Young's modulus E (GPa), Poisson's ratio μ (-), tensile strength YS (MPa), compressive strength 0.23%YS (stress corresponding to 0.23% strain) and compressive strength 0.5%YS (stress corresponding to 0.5% strain)) and pipe forming strain (tensile strain (%)) were input into the generated steel pipe crush strength prediction model to predict the crush strength of the formed steel pipe.

[0171] In addition, in Examples 1 and 2, the actual crushing strength of the steel pipe after forming was measured (actual pipe test results). The judgment criteria for the actual pipe test results in Examples 1 and 2, Examples 3 to 6 shown below, and Comparative Examples 1 to 6 are the same. The difference between the actual crushing strength obtained by the experiment and the standard reference value was evaluated. Steel pipes with actual crushing strength lower than the standard reference value were marked as NG, steel pipes with actual crushing strength higher than the standard reference value within less than 10% were marked as C, steel pipes with actual crushing strength higher than the standard reference value within 10% to less than 20% were marked as B, and steel pipes with actual crushing strength higher than the standard reference value by more than 20% were marked as A. As a result, in Examples 1 and 2, the actual crush strength of the formed steel pipe (actual pipe test results) was lower than the standard reference value (specified standard value), and the judgment result was NG. The predicted value of the crush strength of the formed steel pipe using the steel pipe crush strength prediction model was also lower than the standard reference value (specified standard value), and the judgment result was NG. The experimental evaluation was consistent with the results.

[0172] Furthermore, in Examples 3 to 6, a steel pipe crush strength prediction model is generated using machine learning with multiple training data. These training data include past steel pipe shapes after forming (maximum outer diameter Dmax (mm), minimum outer diameter Dmin (mm), average outer diameter Dave (mm), average plate thickness t (mm), and roundness (Ovality) fO (%)), and past steel pipe strength characteristics after forming (Young's modulus E (GP)). a) The Poisson's ratio μ(-) of the steel pipe, the tensile strength YS (MPa) of the steel pipe, the compressive strength 0.23%YS (stress corresponding to 0.23% strain) and the compressive strength 0.5%YS (stress corresponding to 0.5% strain) of the steel pipe, the pipe forming strain during past steel pipe forming (tensile strain (%) during steel pipe forming), and the coating conditions (maximum temperature (°C) and holding time (min)) are used as input data, and the crushing strength (MPa) of the past coated steel pipe relative to the input data is used as output data.

[0173] Furthermore, in Examples 3 to 6, the generated steel pipe crush strength prediction model was input with the following steel pipe shapes after forming (maximum outer diameter Dmax (mm), minimum outer diameter Dmin (mm), average outer diameter Dave (mm), average plate thickness t (mm), roundness (Ovality) fO (%)), strength characteristics of the formed steel pipe (Young's modulus E (GPa), Poisson's ratio μ (-), tensile strength YS (MPa), compressive strength 0.23%YS (stress corresponding to 0.23% strain), compressive strength 0.5%YS (stress corresponding to 0.5% strain)), pipe forming strain (tensile strain (%)), and coating conditions (maximum temperature (°C) and holding time (min)). The crush strength of the coated steel pipe was then predicted.

[0174] In addition, in Examples 3 to 6, the crushing strength of the coated steel pipe was measured (actual pipe test results).

[0175] As a result, in Examples 3 to 5, the actual crush strength of the coated steel pipe (actual pipe test results) was more than 20% higher than the standard reference value (specified standard value), and the judgment result was A. Furthermore, the predicted crush strength of the coated steel pipe using the steel pipe crush strength prediction model was also more than 20% higher than the standard reference value (specified standard value), and the judgment result was A. The experimental evaluation was consistent with the results. In Example 6, the actual crush strength of the coated steel pipe (actual pipe test results) was higher than the standard reference value (specified standard value) within the range of 10% to less than 20%, and the judgment result was B. Furthermore, the predicted crush strength of the coated steel pipe using the steel pipe crush strength prediction model was also higher than the standard reference value (specified standard value) within the range of 10% to less than 20%, and the judgment result was B. The experimental evaluation was consistent with the results.

[0176] In addition, in Comparative Examples 1 and 2, the steel pipe shape after forming (maximum outer diameter Dmax (mm), minimum outer diameter Dmin (mm), average outer diameter Dave (mm), average plate thickness t (mm), roundness (Ovality) fO (%) of the outer diameter shape of the steel pipe) and the strength characteristics of the steel pipe after forming (Young's modulus E (GPa), Poisson's ratio μ (-) and tensile strength (yield stress corresponding to 0.5% strain)) of the steel pipe were input into the prediction formula shown in Non-Patent Document 1, and the crushing strength of the steel pipe after forming was predicted.

[0177] In addition, in Comparative Examples 3 to 6, the shape of the coated steel pipe after forming (maximum outer diameter Dmax (mm), minimum outer diameter Dmin (mm), average outer diameter Dave (mm), average plate thickness t (mm), roundness fO (%) of the outer diameter shape of the steel pipe) and the strength characteristics of the steel pipe after forming (Young's modulus E (GPa), Poisson's ratio μ (-) and tensile strength (yield stress corresponding to 0.5% strain)) of the steel pipe were input into the prediction formula shown in Non-Patent Document 1, and the crushing strength of the coated steel pipe was predicted.

[0178] As a result, in Comparative Example 1, the actual crush strength of the formed steel pipe (actual pipe test result) was lower than the standard reference value (specified standard value), and the judgment result was NG. However, the predicted value of the crush strength of the formed steel pipe using the prediction formula of Non-Patent Document 1 was higher than the standard reference value (specified standard value) in the range of more than 10% and less than 20%, and the judgment result was B. There was a difference between the two, and the predicted value was inconsistent with the experimental evaluation result.

[0179] In addition, in Comparative Example 2, the actual crush strength of the formed steel pipe (actual pipe test result) was lower than the standard reference value (specified standard value), and the judgment result was NG. However, the predicted value of the crush strength of the formed steel pipe using the prediction formula of Non-Patent Document 1 was higher than the standard reference value (specified standard value) within a range of less than 10%, and the judgment result was C. There was a difference between the two, and the predicted value was inconsistent with the experimental evaluation result.

[0180] In addition, in Comparative Example 3, the actual crush strength of the formed steel pipe (actual pipe test result) was more than 20% higher than the standard reference value (specified standard value), and the judgment result was A. However, the predicted value of the crush strength of the formed steel pipe using the prediction formula of Non-Patent Document 1 was higher than the standard reference value (specified standard value) by less than 10%, and the judgment result was C. There was a difference between the two, and the predicted value was inconsistent with the experimental evaluation result.

[0181] In addition, in Comparative Example 4, the actual crush strength of the coated steel pipe (actual pipe test result) was more than 20% higher than the standard reference value (specified standard value), and the judgment result was A. However, the predicted value of the crush strength of the coated steel pipe using the prediction formula of Non-Patent Document 1 was higher than the standard reference value (specified standard value) in the range of more than 10% and less than 20%, and the judgment result was B. There was a difference between the two, and the predicted value was inconsistent with the experimental evaluation result.

[0182] In addition, in Comparative Example 5, the actual crush strength of the coated steel pipe (actual pipe test result) was more than 20% higher than the specified standard value, and the judgment result was A. However, the predicted value of the crush strength of the coated steel pipe using the prediction formula of Non-Patent Document 1 was higher than the standard reference value (specified standard value) in the range of more than 10% and less than 20%, and the judgment result was B. There was a difference between the two, and the predicted value was inconsistent with the experimental evaluation result.

[0183] In addition, in Comparative Example 6, the actual crush strength of the coated steel pipe (actual pipe test result) was higher than the standard reference value (specified standard value) by less than 10%, and the judgment result was C. However, the predicted value of the crush strength of the coated steel pipe using the prediction formula of Non-Patent Document 1 was higher than the standard reference value (specified standard value) by more than 10% but less than 20%, and the judgment result was B. There was a difference between the two, and the predicted value was inconsistent with the experimental evaluation result.

[0184] Therefore, as illustrated in Examples 1 to 6, in this invention, it was confirmed that the predicted values ​​of the crush strength of the steel pipe after forming and the predicted values ​​of the crush strength of the coated steel pipe were consistent with the experimental results, and the prediction accuracy was high.

[0185] Explanation of reference numerals in the attached figures

[0186] 1… Steel pipe manufacturing characteristics determination device; 2… Calculation device; 3… RAM; 4… ROM; 5… Calculation processing unit; 6… Steel pipe crush strength prediction model generation unit; 7… Steel pipe manufacturing characteristics calculation unit; 8… Input device; 9… Storage device; 10… Output device; 11… Bus; 41… Steel pipe crush strength prediction model generation program; 42… Steel pipe manufacturing characteristics calculation program; 61… Learning data acquisition unit; 62… Preprocessing unit; 63… Model generation unit; 64… Result storage unit; 71… Information reading unit; 72… Crushing strength prediction unit; 73… Steel pipe manufacturing characteristics determination unit; 74… Result output unit; 101… Input layer; 102… Intermediate layer; 103… Output layer.

Claims

1. A method for generating a prediction model for the crushing strength of steel pipes, characterized in that, A steel pipe crush strength prediction model is generated that uses machine learning to predict the crush strength of steel pipes after forming by using multiple learning datasets. The learning data includes past steel pipe manufacturing characteristics, such as the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain obtained through finite element analysis simulation or actual measurement based on the strength characteristics of the steel plate before forming. The output data is the crushing strength of the formed steel pipe relative to the past input data. The shape of the steel pipe after forming includes the maximum outer diameter, the minimum outer diameter, the average outer diameter, the average plate thickness, and the roundness of the outer diameter shape of the steel pipe after forming. The strength characteristics of the steel pipe after forming are: Young's modulus, Poisson's ratio, tensile strength, compressive strength (0.23%YS), and compressive strength (0.5%YS). The tube-forming strain during the forming of the steel pipe is the tensile or compressive strain of the steel plate during the forming process. The 0.23%YS is the stress corresponding to 0.23% strain, and the 0.5%YS is the stress corresponding to 0.5% strain.

2. The method for generating the steel pipe crushing strength prediction model according to claim 1, characterized in that, The machine learning method is a neural network, and the steel pipe crush strength prediction model is a prediction model constructed by a neural network.

3. A method for predicting the crushing strength of steel pipes, characterized in that, The crushing strength prediction model of a steel pipe, generated by the method described in claim 1 or 2, is input with the steel pipe manufacturing characteristics, including the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, and the pipe forming strain, in order to predict the crushing strength of the steel pipe after forming.

4. A method for determining the manufacturing characteristics of a steel pipe, characterized in that, The optimal steel pipe manufacturing characteristics are determined by successively changing at least one of the following: the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, and the pipe-forming strain, which are included in the steel pipe manufacturing characteristics, in a manner that the crush strength of the formed steel pipe predicted by the crush strength prediction method of the steel pipe as described in claim 3 is asymptotically close to the required target crush strength of the formed steel pipe.

5. A method for manufacturing a steel pipe, characterized in that, have: The forming process of steel pipes; The crush strength prediction method for steel pipes as described in claim 3 is used to predict the crush strength of the steel pipe formed in the forming process; and the crush strength prediction process is used to predict the crush strength of the steel pipe formed in the forming process. The crushing strength of the steel pipe predicted by this crushing strength prediction process is associated with the performance prediction value of the steel pipe formed in the forming process and assigned to the process.

6. A method for manufacturing a steel pipe, characterized in that, The manufacturing conditions of the steel pipe are determined based on the optimal manufacturing characteristics of the steel pipe determined by the manufacturing characteristics determination method of the steel pipe according to claim 4, and the steel pipe is manufactured under the determined manufacturing conditions.

7. A method for generating a steel pipe crush strength prediction model, characterized in that, A steel pipe crush strength prediction model is generated by using machine learning to predict the crush strength of coated steel pipes, which are formed by coating after steel pipe forming, based on multiple learning data. The learning data includes past steel pipe manufacturing characteristics, such as the shape of the formed steel pipe, the strength characteristics of the formed steel pipe, the pipe forming strain obtained through finite element analysis simulation based on the strength characteristics of the steel plate before forming, or actual measurement, as well as the coating conditions. The output data is the crushing strength of the coated steel pipe formed after coating in the past, relative to the input data. The shape of the steel pipe after forming includes the maximum outer diameter, the minimum outer diameter, the average outer diameter, the average plate thickness, and the roundness of the outer diameter shape of the steel pipe after forming. The strength characteristics of the steel pipe after forming are: Young's modulus, Poisson's ratio, tensile strength, compressive strength (0.23%YS), and compressive strength (0.5%YS). The tube-forming strain during the forming of the steel pipe is the tensile or compressive strain of the steel plate during the forming process. The 0.23%YS is the stress corresponding to 0.23% strain, and the 0.5%YS is the stress corresponding to 0.5% strain. The coating conditions are the highest temperature and holding time during coating.

8. The method for generating the steel pipe crushing strength prediction model according to claim 7, characterized in that, The machine learning method is a neural network, and the steel pipe crush strength prediction model is a prediction model constructed by a neural network.

9. A method for predicting the crushing strength of a steel pipe, characterized in that, The crush strength of a coated steel pipe, which is the object of prediction, is predicted by inputting the steel pipe manufacturing characteristics, including the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, into the steel pipe crush strength prediction model generated by the method of generating the steel pipe crush strength prediction model according to claim 7 or 8.

10. A method for determining the manufacturing characteristics of a steel pipe, characterized in that, The optimal steel pipe manufacturing characteristics are determined by successively changing at least one of the following: the shape of the steel pipe after forming, the strength characteristics of the steel pipe after forming, the pipe forming strain during forming, and the coating conditions, in a manner that the crush strength of the coated steel pipe predicted by the crush strength prediction method of the steel pipe according to claim 9 is asymptotically close to the target crush strength of the coated steel pipe.

11. A method for manufacturing a steel pipe, characterized in that, have: The process of forming a coated steel pipe by shaping a steel pipe and then coating the formed steel pipe. The crush strength prediction method for predicting the crush strength of a coated steel pipe formed in the coated steel pipe forming process is described in claim 9; and the crush strength prediction process for predicting the crush strength of a coated steel pipe formed in the coated steel pipe forming process; and The crush strength of the coated steel pipe predicted by this crush strength prediction process is associated with the performance prediction value of the coated steel pipe formed in the coated steel pipe forming process and assigned to the process.

12. A method for manufacturing a steel pipe, characterized in that, The manufacturing conditions for the coated steel pipe are determined based on the optimal steel pipe manufacturing characteristics determined by the steel pipe manufacturing characteristic determination method according to claim 10, and the coated steel pipe is manufactured under the determined manufacturing conditions for the coated steel pipe.

Images