WO2021240900A1 - 鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 - Google Patents
鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 Download PDFInfo
- Publication number
- WO2021240900A1 WO2021240900A1 PCT/JP2021/004588 JP2021004588W WO2021240900A1 WO 2021240900 A1 WO2021240900 A1 WO 2021240900A1 JP 2021004588 W JP2021004588 W JP 2021004588W WO 2021240900 A1 WO2021240900 A1 WO 2021240900A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- steel pipe
- forming
- crushing strength
- manufacturing
- prediction model
- Prior art date
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 1401
- 239000010959 steel Substances 0.000 title claims abstract description 1401
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 289
- 238000000034 method Methods 0.000 title claims abstract description 149
- 239000011248 coating agent Substances 0.000 claims abstract description 46
- 238000000576 coating method Methods 0.000 claims abstract description 46
- 238000010801 machine learning Methods 0.000 claims description 18
- 238000013459 approach Methods 0.000 claims description 15
- 238000010422 painting Methods 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 13
- 238000000465 moulding Methods 0.000 abstract description 15
- 238000004364 calculation method Methods 0.000 description 41
- 238000003860 storage Methods 0.000 description 33
- 238000012545 processing Methods 0.000 description 32
- 230000008569 process Effects 0.000 description 30
- 230000035882 stress Effects 0.000 description 21
- 230000006870 function Effects 0.000 description 20
- 238000003062 neural network model Methods 0.000 description 20
- 238000012549 training Methods 0.000 description 13
- 238000012360 testing method Methods 0.000 description 12
- 238000007781 pre-processing Methods 0.000 description 11
- 230000006835 compression Effects 0.000 description 10
- 238000007906 compression Methods 0.000 description 10
- 230000000052 comparative effect Effects 0.000 description 9
- 238000011156 evaluation Methods 0.000 description 7
- 210000002569 neuron Anatomy 0.000 description 5
- 238000005452 bending Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000014759 maintenance of location Effects 0.000 description 2
- 101100328887 Caenorhabditis elegans col-34 gene Proteins 0.000 description 1
- 101150057924 Exoc2 gene Proteins 0.000 description 1
- 102100030843 Exocyst complex component 2 Human genes 0.000 description 1
- 101150110070 SEC5 gene Proteins 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000005536 corrosion prevention Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000008021 deposition Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000003973 paint Substances 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21C—MANUFACTURE OF METAL SHEETS, WIRE, RODS, TUBES OR PROFILES, OTHERWISE THAN BY ROLLING; AUXILIARY OPERATIONS USED IN CONNECTION WITH METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL
- B21C37/00—Manufacture of metal sheets, bars, wire, tubes or like semi-manufactured products, not otherwise provided for; Manufacture of tubes of special shape
- B21C37/06—Manufacture of metal sheets, bars, wire, tubes or like semi-manufactured products, not otherwise provided for; Manufacture of tubes of special shape of tubes or metal hoses; Combined procedures for making tubes, e.g. for making multi-wall tubes
- B21C37/08—Making tubes with welded or soldered seams
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
- G05B13/027—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion using neural networks only
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21C—MANUFACTURE OF METAL SHEETS, WIRE, RODS, TUBES OR PROFILES, OTHERWISE THAN BY ROLLING; AUXILIARY OPERATIONS USED IN CONNECTION WITH METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL
- B21C37/00—Manufacture of metal sheets, bars, wire, tubes or like semi-manufactured products, not otherwise provided for; Manufacture of tubes of special shape
- B21C37/06—Manufacture of metal sheets, bars, wire, tubes or like semi-manufactured products, not otherwise provided for; Manufacture of tubes of special shape of tubes or metal hoses; Combined procedures for making tubes, e.g. for making multi-wall tubes
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N3/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N3/08—Investigating strength properties of solid materials by application of mechanical stress by applying steady tensile or compressive forces
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/41875—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by quality surveillance of production
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2203/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N2203/0014—Type of force applied
- G01N2203/0016—Tensile or compressive
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2203/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N2203/0058—Kind of property studied
- G01N2203/006—Crack, flaws, fracture or rupture
- G01N2203/0067—Fracture or rupture
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2203/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N2203/0058—Kind of property studied
- G01N2203/0069—Fatigue, creep, strain-stress relations or elastic constants
- G01N2203/0075—Strain-stress relations or elastic constants
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2203/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N2203/02—Details not specific for a particular testing method
- G01N2203/0202—Control of the test
- G01N2203/0212—Theories, calculations
- G01N2203/0216—Finite elements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2203/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N2203/02—Details not specific for a particular testing method
- G01N2203/026—Specifications of the specimen
- G01N2203/0262—Shape of the specimen
- G01N2203/0274—Tubular or ring-shaped specimens
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2203/00—Investigating strength properties of solid materials by application of mechanical stress
- G01N2203/02—Details not specific for a particular testing method
- G01N2203/026—Specifications of the specimen
- G01N2203/0298—Manufacturing or preparing specimens
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N3/00—Investigating strength properties of solid materials by application of mechanical stress
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32193—Ann, neural base quality management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Definitions
- the present invention relates to a method for generating a steel pipe crushing strength prediction model, a method for predicting the crushing strength of a steel pipe, a method for determining a steel pipe manufacturing characteristic, and a method for manufacturing a steel pipe.
- crushing also called collapse
- a steel pipe line pipe
- a steel pipe having excellent pressure-resistant crushing performance as a steel pipe for use such as a submarine pipeline in which a high compressive stress acts.
- Non-Patent Document 1 a method for predicting and evaluating the pressure-resistant crushing performance of the target steel pipe described in Non-Patent Document 1 is known.
- standards such as DNV-F01 are established as a method for predicting and evaluating the pressure resistance crushing performance of the target steel pipe, and the Ovality of the outer diameter shape and the material wall thickness of the evaluation target steel pipe (steel pipe after steel pipe forming) are established.
- Non-Patent Document 1 has the following problems.
- the crushing strength of the steel pipe is not only the shape of the steel pipe after forming the steel pipe and the strength characteristics of the steel pipe after forming the steel pipe (tensile strength, compressive strength, Young's ratio, Poisson's ratio, etc.), but also the pipe forming strain during forming the steel pipe (Poreson's ratio, etc.). It also depends on the strain history during steel pipe forming). This is because the pipe forming strain during steel pipe forming has a great influence on the shape of the steel pipe after forming the steel pipe and the strength characteristics of the steel pipe after forming the steel pipe.
- Non-Patent Document 1 the pipe forming strain during steel pipe forming is not taken into consideration, and the accuracy of the predicted value of the crushing strength of the steel pipe is poor. Does not match, and the difference is large. For this reason, when designing a steel pipe, the design may be excessively safe, or crushing may occur at an external pressure lower than expected, resulting in a major accident.
- the present invention has been made to solve this conventional problem, and an object thereof is to paint a steel pipe after forming a steel pipe or a coating after forming a steel pipe in consideration of pipe forming strain at the time of forming a steel pipe. It is an object of the present invention to provide a method for generating a steel pipe crushing strength prediction model, a method for predicting a steel pipe crushing strength, a method for determining a steel pipe manufacturing characteristic, and a method for manufacturing a steel pipe, which can predict the crushing strength of a steel pipe with high accuracy.
- the method for generating a steel pipe crushing strength prediction model includes a steel pipe shape after steel pipe forming, a steel pipe strength characteristic after steel pipe forming, and a pipe forming strain during steel pipe forming. Predicting the crushing strength of steel pipes after forming steel pipes by machine-learning multiple learning data using the past steel pipe manufacturing characteristics as input data and the crushing strength of steel pipes after forming past steel pipes as output data for this input data.
- the gist is to generate a steel pipe crushing strength prediction model.
- the method for predicting the crushing strength of a steel pipe is based on the steel pipe crushing strength prediction model generated by the above-mentioned method for generating the steel pipe crushing strength prediction model, and the steel pipe after forming the steel pipe to be predicted.
- the gist is to predict the crushing strength of the steel pipe after steel pipe forming by inputting the shape, the steel pipe strength characteristics after steel pipe forming, and the steel pipe manufacturing characteristics including the pipe forming strain during steel pipe forming.
- the crushing strength of the steel pipe after forming the steel pipe predicted by the above-mentioned method for predicting the crushing strength of the steel pipe is required to be the target steel pipe after forming the steel pipe.
- Optimal by sequentially changing at least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the pipe forming strain during steel pipe forming so as to gradually approach the crushing strength of The gist is to determine the manufacturing characteristics of steel pipes.
- the crushing strength of the steel pipe formed in the forming step is predicted by the forming step of the steel pipe for forming the steel pipe and the above-mentioned crushing strength prediction method for the steel pipe.
- the gist is to include a crushing strength prediction step and a performance prediction value assigning step of associating the crushing strength of the steel pipe predicted by the crushing strength prediction step with the steel pipe formed in the forming step.
- the manufacturing conditions for the steel pipe are determined based on the optimum steel pipe manufacturing characteristics determined by the above-mentioned method for determining the manufacturing characteristics of the steel pipe, and the determined steel pipe is used.
- the gist is to manufacture steel pipes under manufacturing conditions.
- a method for generating a steel pipe crushing strength prediction model includes past steel pipe shapes after steel pipe forming, steel pipe strength characteristics after steel pipe forming, pipe forming strain during steel pipe forming, and coating conditions. Paint made by painting after steel pipe forming by machine-learning multiple learning data using the crushing strength of the painted steel pipe as output data, using the steel pipe manufacturing characteristics as input data and painting after past steel pipe forming for this input data. The gist is to generate a steel pipe crushing strength prediction model that predicts the crushing strength of steel pipes.
- the method for predicting the crushing strength of a steel pipe is based on the steel pipe crushing strength prediction model generated by the above-mentioned method for generating the steel pipe crushing strength prediction model, after forming the steel pipe of the coated steel pipe to be predicted.
- the gist is to predict the crushing strength of coated steel pipes that are painted after steel pipe molding by inputting the steel pipe shape, steel pipe strength characteristics after steel pipe forming, pipe forming strain during steel pipe forming, and steel pipe manufacturing characteristics including coating conditions. do.
- the crushing strength of the coated steel pipe predicted by the above-mentioned crushing strength prediction method of the steel pipe gradually approaches the required target crushing strength of the coated steel pipe.
- at least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions included in the steel pipe manufacturing characteristics are sequentially changed to produce the optimum steel pipe.
- the gist is to determine the characteristics.
- the method for manufacturing a steel pipe according to another aspect of the present invention is based on a coated steel pipe forming step of forming a steel pipe and painting the formed steel pipe to form a coated steel pipe, and the above-mentioned method for predicting the crushing strength of the steel pipe.
- the crushing strength prediction step for predicting the crushing strength of the coated steel pipe formed in the coated steel pipe forming step and the crushing strength of the coated steel pipe predicted by the crushing strength prediction step are applied to the coated steel pipe formed in the coated steel pipe forming step.
- the gist is to have a performance prediction value assigning process to be linked.
- the manufacturing conditions for the coated steel pipe are determined based on the optimum steel pipe manufacturing characteristics determined by the above-mentioned method for determining the manufacturing characteristics of the steel pipe, and the determined coating is performed.
- the gist is to manufacture coated steel pipes under the manufacturing conditions of steel pipes.
- the method for generating the steel pipe crushing strength prediction model the method for predicting the crushing 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 present invention, after the steel pipe is formed in consideration of the pipe forming strain during the steel pipe forming.
- a manufacturing method can be provided.
- FIG. 1 shows a schematic configuration of a steel pipe manufacturing characteristic determination device to which a method for generating a steel pipe crushing strength prediction model, a method for predicting crushing strength of a steel pipe, and a method for determining a steel pipe manufacturing characteristic according to the first embodiment of the present invention are applied.
- a functional block diagram is shown.
- the steel pipe manufacturing characteristic determination device 1 applied to the first embodiment shown in FIG. 1 generates a steel pipe crushing strength prediction model and predicts the crushing strength of the steel pipe after steel pipe forming using the generated steel pipe crushing strength prediction model. I do. Further, the steel pipe manufacturing characteristic determining device 1 determines the optimum steel pipe manufacturing characteristics so that the predicted crushing strength of the steel pipe after forming the steel pipe is gradually close to the required crushing strength of the steel pipe after forming the steel pipe.
- the steel pipe manufacturing characteristic determination device 1 shown in FIG. 1 is a computer system including an arithmetic unit 2, an input device 8, a storage device 9, and an output device 10.
- the arithmetic unit 2 includes a RAM 3, a ROM 4, and an arithmetic processing unit 5.
- the RAM 3, the ROM 4, the arithmetic processing unit 5, the input device 8, the storage device 9, and the output device 10 are connected by a bus 11.
- the arithmetic unit 2 and the input device 8, the storage device 9, and the output device 10 are not limited to this connection mode, and may be connected wirelessly or may be connected by a combination of wired and wireless devices. good.
- the input device 8 functions as an input port into which various information is input by the operator of the system, such as a keyboard, a pen tablet, a touch pad, and a mouse.
- the input device 8 has, for example, a command for generating a steel pipe crushing strength prediction model, a calculation command for steel pipe manufacturing characteristics, a steel pipe shape after steel pipe forming of a steel pipe to be predicted for crushing strength, a steel pipe strength characteristic after steel pipe forming, and a steel pipe.
- Steel pipe manufacturing characteristics consisting of pipe forming strain during forming, crushing strength of steel pipe after target steel pipe forming, and steel pipe manufacturing characteristic determination mode information are input.
- the steel pipe is generally manufactured by bending a plate-shaped steel plate into a circular pipe shape, and then coating the surface thereof.
- the steel pipe shape after forming the steel pipe among the steel pipe manufacturing characteristics input to the input device 8 means the shape of the steel pipe after forming the steel plate into a circular pipe shape.
- the shape of the steel pipe after forming the steel pipe includes the maximum outer diameter Dmax (mm) of the steel pipe, the minimum outer diameter Dmin (mm) of the steel pipe, the average outer diameter Dave (mm) of the steel pipe, and the average plate thickness t of the steel pipe (. mm), and the roundness (Ovality) fO (%) of the outer diameter shape of the steel pipe.
- the measured one is input to the input device 8.
- the crushing strength of the steel pipe means the load stress (MPa) when the steel pipe is crushed, and the "crush” here means the maximum value of the load stress and can maintain its shape against external pressure. It shall be in a deformed state until it disappears.
- the steel pipe strength characteristic after forming the steel pipe means the strength characteristic of the steel pipe after forming the steel plate into a pipe shape.
- the steel pipe strength characteristics after forming the steel pipe include Young's modulus E (GPa) of the steel pipe, Poisson's ratio ⁇ (-) of the steel pipe, tensile strength YS (MPa) of the steel pipe, and compressive strength 0.23% YS of the steel pipe. (Stress corresponding to 0.23% strain) and compressive strength of steel pipe 0.5% YS (stress corresponding to 0.5% strain). Since the steel pipe strength characteristic after forming the steel pipe has a great influence on the predicted crushing strength of the steel pipe after forming the steel pipe, it is input. As the steel pipe strength characteristics after steel pipe forming, those obtained by simulating from the strength characteristics of the steel sheet before steel pipe forming by finite element analysis and those actually measured are input.
- the pipe forming strain during steel pipe forming is a tensile strain (%) or a compressive strain (%) during steel pipe forming.
- the pipe forming strain during steel pipe forming has a great influence on the steel pipe shape after steel pipe forming and the steel pipe strength characteristics after steel pipe forming, and as a result, has a great influence on the predicted crushing strength of the steel pipe after steel pipe forming. Therefore, I tried to enter it.
- the pipe forming strain during steel pipe forming the one obtained by forming simulation by finite element analysis from the strength characteristics of the steel sheet before steel pipe forming or the measured one is input.
- the storage device 9 is composed of, for example, a hard disk drive, a semiconductor drive, an optical drive, etc., and realizes the functions required in this system (the steel pipe crushing strength prediction model generation unit 6 and the steel pipe manufacturing characteristic calculation unit 7, which will be described later). It is a device that stores information necessary for.
- the information necessary for realizing the function by the steel pipe crush strength prediction model generation unit 6 includes, for example, the past including the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming.
- an input device input to the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6 and the steel pipe crush strength prediction model.
- Steel pipe shape after steel pipe forming of steel pipe to be predicted of crushing strength input in 8 steel pipe strength characteristics after steel pipe forming, steel pipe manufacturing characteristics consisting of pipe forming strain during steel pipe forming, steel pipe after target steel pipe forming Crushing strength and steel pipe manufacturing characteristic determination mode information (information on whether or not the mode determines the optimum steel pipe manufacturing characteristic) can be mentioned.
- the output device 10 is used to output data from the arithmetic unit 2, for example, information on the crushing strength (predicted value) of the steel pipe after forming the steel pipe predicted by the crushing strength prediction unit 72, which will be described later, and the steel pipe manufacturing characteristic determination unit 73. It functions as an output port that outputs information on the determined optimum steel pipe manufacturing characteristics.
- the output device 10 can display a screen based on the output data by including an arbitrary display such as a liquid crystal display or an organic display, for example.
- the arithmetic unit 2 includes a RAM 3, a ROM 4, and an arithmetic processing unit 5.
- the ROM 4 stores a steel pipe crush strength prediction model generation program 41 and a steel pipe manufacturing characteristic calculation program 42.
- the arithmetic processing unit 5 has an arithmetic processing function and is connected to the RAM 3 and the ROM 4 by a bus 11. Further, the RAM 3, the ROM 4, and the arithmetic processing unit 5 are connected to the input device 8, the storage device 9, and the output device 10 via the bus 11.
- the calculation processing unit 5 includes a steel pipe crush strength prediction model generation unit 6 and a steel pipe manufacturing characteristic calculation unit 7 as functional blocks.
- the steel pipe crushing strength prediction model generation unit 6 of the arithmetic processing unit 5 is a past unit composed of the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming stored in the storage device 9.
- a steel pipe crushing strength prediction model is generated by machine-learning a plurality of training data using the steel pipe manufacturing characteristics as input data and the crushing strength of the steel pipe after past steel pipe forming as output data for this input data.
- the machine learning method is a neural network
- the steel pipe crush strength prediction model is a prediction model constructed by a neural network.
- the steel pipe crush strength prediction model generation unit 6 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. Then, when the steel pipe crush strength prediction model generation unit 6 inputs the generation command of the steel pipe crush strength prediction model to the input device 8 and receives the generation command of the steel pipe crush strength prediction model, the steel pipe crush stored in the ROM 4 is received.
- the intensity prediction model generation program 41 is executed, and each function of the training data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64 is executed.
- the steel pipe crush strength prediction model is updated every time the steel pipe crush strength prediction model generation unit 6 executes each function.
- Execution processing of each function of the learning data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64 by the steel pipe crush strength prediction model generation unit 6 is the steel pipe according to the first embodiment of the present invention.
- Input data is past steel pipe manufacturing characteristics consisting of steel pipe shape after forming, steel pipe strength characteristics after steel pipe forming, and pipe forming strain during steel pipe forming, and the crushing strength of steel pipes after past steel pipe forming with respect to this input data is output. It corresponds to a method of generating a steel pipe crushing strength prediction model that generates a steel pipe crushing strength prediction model that predicts the crushing strength of a steel pipe after steel pipe forming by machine learning a plurality of training data as data.
- the learning data acquisition unit 61 obtains the past steel pipe manufacturing characteristics stored in the storage device 9, which are the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming.
- As input data a process is performed to acquire a plurality of training data using the crushing strength of the steel pipe after forming the past steel pipe as output data for this input data.
- Each learning data consists of a set of input data and output data.
- the preprocessing unit 62 processes a plurality of learning data acquired by the learning data acquisition unit 61 for generating a steel pipe crush strength prediction model.
- the pretreatment unit 62 provides actual information on past steel pipe manufacturing characteristics including the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming, which constitute the learning data. Is standardized (normalized) between 0 and 1 in order to be read into the neural network model.
- the model generation unit 63 machine-learns a plurality of learning data preprocessed by the pretreatment unit 62, and the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming.
- a process is performed to generate a steel pipe crushing strength prediction model that includes the past steel pipe manufacturing characteristics consisting of the above as input data and the crushing strength of the steel pipe after past steel pipe forming as output data.
- a neural network is adopted as a machine learning method, a neural network model is generated as a steel pipe crush strength prediction model.
- the model generation unit 63 has input actual data (past actual data of steel pipe manufacturing characteristics) and output actual data (crushing strength of the steel pipe after steel pipe forming) in the training data processed for generating the steel pipe crushing strength prediction model.
- the neural network model is expressed, for example, by a functional expression.
- the model generation unit 63 sets the hyperparameters used in the neural network model, and also performs learning by the neural network model using those hyperparameters.
- hyperparameters 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 are usually set, but are not limited to this.
- FIG. 2 shows a processing flow of a steel pipe crushing strength prediction model, which is a neural network model generated by the method for generating a steel pipe crushing strength prediction model according to the first embodiment of the present invention.
- the steel pipe crush strength prediction model which is a neural network model, includes an input layer 101, an intermediate layer 102, and an output layer 103 in this order from the input side.
- the model generation unit 63 performs training by the neural network model using the hyper parameter
- the input layer 101 has a steel pipe shape after steel pipe forming and a steel pipe shape after steel pipe forming which constitutes the learning data processed by the preprocessing unit 62.
- the actual information of the past steel pipe manufacturing characteristics consisting of the steel pipe strength characteristics and the pipe forming strain during steel pipe forming, that is, the actual information of the past steel pipe manufacturing characteristics standardized between 0 and 1, is stored.
- the intermediate layer 102 is composed of a plurality of hidden layers, and a plurality of neurons are arranged in each hidden layer.
- the number of hidden layers formed in the intermediate layer 102 is not particularly limited, but empirically, if there are too many hidden layers, the prediction accuracy will decrease, so that the number of hidden layers is preferably 5 or less.
- the output layer 103 is combined with the neuron information transmitted by the intermediate layer 102, and is output as the crushing strength of the steel pipe after the final steel pipe molding. Learning is performed by gradually optimizing the weighting coefficient in the neural network model based on the output result and the actual crushing strength of the steel pipe after the past steel pipe forming that has been read.
- the result storage unit 64 stores the training data, the parameters (weighting factors) of the neural network model, and the output result of the neural network model for the training data in the storage device 9.
- the steel pipe manufacturing characteristic calculation unit 7 of the calculation processing unit 5 uses the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6 to determine the steel pipe shape and steel pipe forming of the steel pipe whose crush strength is to be predicted.
- a process of predicting the crushing strength of the steel pipe after steel pipe forming corresponding to the steel pipe manufacturing characteristic is performed by inputting the steel pipe strength characteristic after that and the steel pipe manufacturing characteristic consisting of the pipe making strain at the time of steel pipe forming.
- the steel pipe manufacturing characteristic determination mode information is the steel pipe manufacturing characteristic determination mode
- the steel pipe manufacturing characteristic calculation unit 7 determines the predicted crushing strength of the steel pipe after steel pipe forming, which is the target after steel pipe forming.
- Optimal by sequentially changing at least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the pipe forming strain during steel pipe forming so as to gradually approach the crushing strength of Performs processing to determine steel pipe manufacturing characteristics.
- the steel pipe manufacturing characteristic calculation unit 7 includes 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 as functional blocks. ing.
- the information reading unit 71 performs a process of reading information necessary for realizing the function of the steel pipe manufacturing characteristic calculation unit 7 stored in the storage device 9. Specifically, the information reading unit 71 performs a process of reading the steel pipe crushing strength prediction model generated by the steel pipe crushing strength prediction model generation unit 6. Further, the information reading unit 71 is based on the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe making strain during steel pipe forming, which are input to the steel pipe crushing strength prediction model. Information on the steel pipe manufacturing characteristics, information on the crushing strength of the steel pipe after forming the target steel pipe, and information on the steel pipe manufacturing characteristic determination mode are read.
- the crushing strength prediction unit 72 includes a steel pipe shape after steel pipe forming, a steel pipe strength characteristic after steel pipe forming, and a pipe forming strain during steel pipe forming, which are read by the information reading unit 71 and are to be predicted by the information reading unit 71.
- the steel pipe manufacturing characteristics are input to the steel pipe crushing strength prediction model read by the information reading unit 71, and a process of predicting the crushing strength of the steel pipe after forming the steel pipe is performed.
- the steel pipe manufacturing characteristic determination unit 73 and the crushing strength prediction unit 72 crush the steel pipe after the predicted steel pipe forming when the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is the steel pipe manufacturing characteristic determination mode.
- the steel pipe manufacturing characteristic determination unit 73 predicts the crushing strength of the steel pipe after the steel pipe is formed by the crushing strength prediction unit 72.
- Information (predicted value) is output to the result output unit 74.
- the result output unit 74 performs a process of outputting the determined optimum steel pipe manufacturing characteristic information or the predicted steel pipe crushing strength information (predicted value) after forming the steel pipe to the output device 10, and these A process of storing information in the storage device 9 is performed.
- the steel pipe manufacturing characteristic calculation unit 7 inputs a steel pipe manufacturing characteristic calculation command to the input device 8 and receives the steel pipe manufacturing characteristic calculation command
- the steel pipe manufacturing characteristic calculation unit 7 executes the steel pipe manufacturing characteristic calculation program 42 stored in the ROM 4 and performs information.
- Each function of the reading unit 71, the crush strength prediction unit 72, the steel pipe manufacturing characteristic determination unit 73, and the result output unit 74 is executed.
- the information reading unit 71 of the steel pipe manufacturing characteristic calculation unit 7 reads the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6 stored in the storage device 9 in step S1.
- the information reading unit 71 reads information on the crushing strength of the steel pipe after forming the required target steel pipe, which is input from a higher-level computer (not shown) and stored in the storage device 9.
- step S3 the information reading unit 71 is input to the input device 8 by the operator, and is input to the steel pipe crushing strength prediction model stored in the storage device 9 after the steel pipe is formed to be the prediction target of the crushing strength.
- Information on steel pipe shape, steel pipe strength characteristics after steel pipe forming, and steel pipe manufacturing characteristics consisting of pipe forming strain during steel pipe forming is read.
- step S4 the information reading unit 71 inputs to the input device 8 by the operator and stores the steel pipe manufacturing characteristic determination mode information (information on whether or not the mode determines the optimum steel pipe manufacturing characteristic) stored in the storage device 9. ) Is read.
- step S5 the crushing strength prediction unit 72 applies the steel pipe crushing strength prediction model read in step S1 to the steel pipe shape and steel pipe forming after steel pipe forming of the steel pipe to be predicted in the crushing strength read in step S3.
- the crushing strength of the steel pipe after steel pipe forming is predicted.
- steps S1 to S5 the steel pipe shape after forming the steel pipe to be predicted is applied to the steel pipe crush strength prediction model generated by the method for generating the steel pipe crush strength prediction model according to the first embodiment of the present invention.
- the method of predicting the crushing strength of steel pipes by inputting the steel pipe strength characteristics after steel pipe forming and the steel pipe manufacturing characteristics consisting of the pipe forming strain during steel pipe forming to predict the crushing strength of the steel pipe after steel pipe forming.
- step S6 in step S6, the steel pipe manufacturing characteristic determination mode information (information on whether or not the mode is the mode for determining the optimum steel pipe manufacturing characteristic) read in step S4 is the steel pipe manufacturing characteristic determination mode (optimal). It is determined whether or not the mode is used to determine the manufacturing characteristics of steel pipes.
- the determination result in step S6 is YES (in the steel pipe manufacturing characteristic determination mode)
- the process proceeds to step S7, and when the determination result in step S6 is NO (not in the steel pipe manufacturing characteristic determination mode), step S9. Move to.
- step S7 the steel pipe manufacturing characteristic determination unit 73 determines the crushing strength (predicted value) of the steel pipe after steel pipe forming predicted in step S5, and the required target steel pipe after steel pipe forming read in step S2. It is determined whether or not the difference from the crushing strength (target value) is within a predetermined threshold.
- the above-mentioned predetermined threshold value is set to about 0.5% to 1%, although it varies depending on the target value and the manufacturing conditions.
- step S7 When 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 predetermined threshold value), the process proceeds to step S8, and when the determination result in step S7 is NO (predicted value). When it is determined that the difference between the target value and the target value is larger than the predetermined threshold value), the process proceeds to step S10.
- step S10 the steel pipe manufacturing characteristic determination unit 73 determines the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the steel pipe forming time in the steel pipe manufacturing characteristic of the steel pipe to be predicted of the crushing strength read in step S3. Change at least one of the pipe forming strains of the above and return to step S5.
- the crushing strength prediction unit 72 changed at least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming in step S10.
- the steel pipe manufacturing characteristics of the steel pipe are input to the steel pipe crushing strength prediction model read in step S1, and the crushing strength of the steel pipe after steel pipe forming is predicted again.
- the steel pipe manufacturing characteristic determination unit 73 is requested to read the crushing strength (predicted value) of the steel pipe after forming the steel pipe predicted again in step S5 in step S7 and the crushing strength (predicted value) of the steel pipe in step S2.
- step S10 It is determined whether or not the difference from the crushing strength (target value) of the steel pipe after forming the target steel pipe is within a predetermined threshold. Then, a series of steps of step S10, step S5, step S6, and step S7 are repeatedly executed until the determination result becomes YES.
- step S7 when 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 predetermined threshold value), the process proceeds to step S8.
- step S8 the steel pipe manufacturing characteristic determination unit 73 determines that the difference between the predicted value and the target value is within a predetermined threshold value, the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the steel pipe forming.
- the steel pipe manufacturing characteristics consisting of the pipe making strain at the time are determined as the optimum steel pipe manufacturing characteristics.
- step S9 the result output unit 74 of the steel pipe manufacturing characteristic calculation unit 7 determines the optimum steel pipe manufacturing characteristic determined in step S8 when the determination result in step S6 is YES (in the steel pipe manufacturing characteristic determination mode). Information is output to the output device 10.
- the result output unit 74 outputs information (predicted value) of the crushing strength of the steel pipe after forming the steel pipe predicted in step S5. Output to device 10. As a result, the processing of the steel pipe manufacturing characteristic calculation unit 7 is completed.
- the method for generating the steel pipe crushing strength prediction model according to the first embodiment of the present invention includes the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming in the past.
- Generate a crush strength prediction model (steel pipe crush strength prediction model generation unit 6). This makes it possible to appropriately generate a steel pipe crushing strength prediction model for predicting the crushing strength of a steel pipe after steel pipe forming with high accuracy in consideration of the pipe forming strain during steel pipe forming.
- the method for predicting the crushing strength of the steel pipe according to the first embodiment of the present invention is based on the steel pipe crushing strength prediction model generated by the method for generating the steel pipe crushing strength prediction model, and the shape of the steel pipe after forming the steel pipe to be predicted.
- Steel pipe strength characteristics after steel pipe forming and steel pipe manufacturing characteristics consisting of pipe forming strain during steel pipe forming are input to predict the crushing strength of the steel pipe after steel pipe forming (steps S1 to S5).
- the crushing strength of the steel pipe after forming the steel pipe can be predicted with high accuracy in consideration of the pipe forming strain during the forming of the steel pipe.
- the predicted crushing strength of the steel pipe after forming the steel pipe is made to gradually approach the required target crushing strength of the steel pipe after forming the steel pipe.
- At least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the pipe forming strain during steel pipe forming, which are included in the steel pipe manufacturing characteristics, is sequentially changed to determine the optimum steel pipe manufacturing characteristics (. Step S6, Step S7, Step S10, Step S5, Step S6, Step S7 and Step S8).
- the steel pipe formed in the forming step by the method for forming the steel pipe and the method for predicting the crushing strength of the steel pipe (steps S1 to S5). It may be provided with a crushing strength prediction step of predicting the crushing strength and a performance prediction value assigning step of associating the crushing strength of the steel pipe predicted by the crushing strength prediction step with the steel pipe formed in the forming step.
- the predicted crushing strength (predicted value) of the steel pipe is given to the formed steel pipe by marking. Or, it is achieved by attaching a tag stating the predicted crushing strength (predicted value) of the predicted steel pipe to the formed steel pipe. As a result, the person who handles the formed steel pipe can grasp the crushing strength (predicted value) of the steel pipe.
- the manufacturing conditions of the steel pipe selection of the pipe making method, bending ratio at the time of pipe making, manufacturing
- the steel pipe may be manufactured under the determined manufacturing conditions of the steel pipe.
- the steel pipe manufacturing method according to the first embodiment of the present invention is the optimum method determined by the steel pipe manufacturing characteristic determination method (step S6, step S7, step S10, step S5, step S6, step S7 and step S8).
- the manufacturing conditions of the steel pipe may be determined based on the manufacturing characteristics of the steel pipe, and the steel pipe may be manufactured under the determined manufacturing conditions of the steel pipe.
- the manufactured steel pipe satisfies the determined optimum steel pipe manufacturing characteristics, and as a result, the predicted crushing strength (predicted value) of the steel pipe becomes the required crushing strength of the steel pipe after forming the steel pipe. It becomes a gradual one, and it becomes a steel pipe with excellent pressure-resistant crushing performance, and it is possible to avoid damage or damage accident of the structure.
- FIGS. 1 and 4 to 5 For a method of generating a steel pipe crushing strength prediction model, a method of predicting a steel pipe crushing strength, a method of determining a steel pipe manufacturing characteristic, and a method of manufacturing a steel pipe according to the second embodiment of the present invention. I will explain. The description of the members already described in the first embodiment may be omitted.
- the steel pipe manufacturing characteristic determining device 1 shown in FIG. 1 is also applied to the method for generating a steel pipe crushing strength prediction model, the method for predicting the crushing strength of a steel pipe, and the method for determining a steel pipe manufacturing characteristic according to the second embodiment.
- the method for generating a steel pipe crushing strength prediction model according to the second embodiment is to generate a steel pipe crushing strength prediction model for a coated steel pipe that is painted after forming the steel pipe.
- the method for predicting the crushing strength of a steel pipe according to the second embodiment predicts the crushing strength of a coated steel pipe that is painted after forming the steel pipe by using the generated steel pipe crushing strength prediction model.
- the optimum crushing strength of the coated steel pipe is determined so that the predicted crushing strength of the coated steel pipe approaches the required target crushing strength of the coated steel pipe.
- the steel pipe manufacturing characteristic determination device 1 is a computer system including an arithmetic unit 2, an input device 8, a storage device 9, and an output device 10, and the basic configuration thereof is already present. Since it has been explained, the description will be omitted as appropriate. Similar to the first embodiment, the input device 8 is input with a command for generating a steel pipe crushing strength prediction model, a command for calculating steel pipe manufacturing characteristics, and the like. In the second embodiment, unlike the first embodiment, the crushing strength of the coated steel pipe coated after forming the steel pipe is predicted. Therefore, as a steel pipe manufacturing characteristic, the steel pipe after forming the steel pipe of the coated steel pipe to which the crushing strength is predicted is predicted. In addition to the shape, steel pipe strength characteristics after steel pipe forming, and pipe forming strain during steel pipe forming, coating conditions are input. Further, the crushing strength of the coated steel pipe to be painted after the target steel pipe shape is input to the input device 8.
- the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming are the same as those in the first embodiment, but the coating conditions are the maximum temperature (° C.) at the time of coating. And the holding time (min). As for this painting condition, the measured one is input. Paint is applied to the formed steel pipe for corrosion prevention, and in particular, the steel pipe used in the submarine pipeline has excellent corrosion resistance, so that it is generally applied after molding.
- the coating conditions (maximum temperature (° C.) and holding time (min)) in this coating affect the strength characteristics of the steel pipe after forming the steel pipe and directly affect the crushing performance of the coated steel pipe. bottom. Due to the influence of coating heating of the coating, the material of the steel pipe changes (deposition of dislocations, recovery, strain aging, etc.), and the crushing strength increases from the crushing strength of the steel pipe after forming the steel pipe (crushing performance before painting). descend.
- the storage device 9 is a device that stores information necessary for realizing the functions of the steel pipe crush strength prediction model generation unit 6 and the steel pipe manufacturing characteristic calculation unit 7.
- Information necessary for realizing the function by the steel pipe crush strength prediction model generation unit 6 includes past steel pipe manufacturing consisting of steel pipe shape after steel pipe forming, steel pipe strength characteristics after steel pipe forming, pipe forming strain during steel pipe forming, and coating conditions.
- a plurality of learning data can be mentioned in which the characteristic is used as input data and the crushing strength of the coated steel pipe painted after the past steel pipe forming with respect to this input data is used as output data.
- the information necessary for realizing the function by the steel pipe manufacturing characteristic calculation unit 7 is input to the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6 and the input device 8 input to the steel pipe crush strength prediction model.
- the crushing strength of the coated steel pipe and the steel pipe manufacturing characteristic determination mode information (information on whether or not the mode determines the optimum steel pipe manufacturing characteristic) can be mentioned.
- the output device 10 includes output data from the arithmetic unit 2, for example, information on the crushing strength (predicted value) of the coated steel pipe that is painted after forming the steel pipe predicted by the crushing strength prediction unit 72, and a steel pipe manufacturing characteristic determination unit. It functions as an output port that outputs information on the optimum steel pipe manufacturing characteristics determined in 73.
- the arithmetic unit 2 has the same configuration as that of the first embodiment, and in particular, includes a steel pipe crush strength prediction model generation unit 6 and a steel pipe manufacturing characteristic calculation unit 7 as functional blocks.
- the steel pipe crush strength prediction model generation unit 6 of the arithmetic processing unit 5 is stored in the storage device 9 from the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions.
- a steel pipe crushing strength prediction model by machine-learning multiple learning data using the past steel pipe manufacturing characteristics as input data and the crushing strength of the coated steel pipe painted after forming the past steel pipe as output data for this input data.
- the machine learning method is a neural network as in the first embodiment, and the steel pipe crush strength prediction model is a prediction model constructed by the neural network.
- the steel pipe crush strength prediction model generation unit 6 has the learning data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64 as functional blocks, as in the first embodiment. It is equipped with. Then, when the steel pipe crush strength prediction model generation unit 6 inputs the generation command of the steel pipe crush strength prediction model to the input device 8 and receives the generation command of the steel pipe crush strength prediction model, the steel pipe crush stored in the ROM 4 is received.
- the intensity prediction model generation program 41 is executed, and each function of the training data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64 is executed.
- the steel pipe crush strength prediction model is updated every time the steel pipe crush strength prediction model generation unit 6 executes each function.
- Execution processing of each function of the learning data acquisition unit 61, the preprocessing unit 62, the model generation unit 63, and the result storage unit 64 by the steel pipe crush strength prediction model generation unit 6 is the steel pipe according to the second embodiment of the present invention.
- the input data is the past steel pipe manufacturing characteristics consisting of the steel pipe shape after forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the painting conditions, and the input data is painted after the past steel pipe forming.
- a steel pipe crushing strength prediction model that generates a steel pipe crushing strength prediction model that predicts the crushing strength of a coated steel pipe that is painted after steel pipe molding by machine learning multiple learning data with the crushing strength of the coated steel pipe as output data. Corresponds to the generation method.
- the learning data acquisition unit 61 is a past steel pipe stored in the storage device 9, which includes the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions.
- a process is performed in which a plurality of learning data are acquired using the manufacturing characteristics as input data and the crushing strength of the coated steel pipe painted after the past steel pipe forming for this input data as output data.
- Each learning data consists of a set of input data and output data.
- the preprocessing unit 62 processes a plurality of learning data acquired by the learning data acquisition unit 61 for generating a steel pipe crush strength prediction model, as in the first embodiment.
- the model generation unit 63 machine-learns a plurality of learning data preprocessed by the pretreatment unit 62 to obtain a steel pipe shape after steel pipe forming, a steel pipe strength characteristic after steel pipe forming, a pipe forming strain during steel pipe forming, and the like. And, a process is performed to generate a steel pipe crushing strength prediction model that includes the past steel pipe manufacturing characteristics consisting of coating conditions as input data and outputs the crushing strength of the coated steel pipe that is coated after the past steel pipe forming.
- a neural network model is generated as a steel pipe crush strength prediction model.
- FIG. 4 shows a processing flow of a steel pipe crushing strength prediction model, which is a neural network model generated by the method for generating a steel pipe crushing strength prediction model according to the second embodiment of the present invention.
- the steel pipe crush strength prediction model which is a neural network model, includes an input layer 101, an intermediate layer 102, and an output layer 103 in this order from the input side.
- the model generation unit 63 performs training by the neural network model using the hyper parameter
- the input layer 101 has a steel pipe shape after steel pipe forming and a steel pipe shape after steel pipe forming which constitutes the learning data processed by the preprocessing unit 62.
- the actual information of the past steel pipe manufacturing characteristics including the steel pipe strength characteristics, the pipe forming strain during steel pipe forming, and the coating conditions, that is, the actual information of the past steel pipe manufacturing characteristics standardized between 0 and 1 is stored. ..
- the intermediate layer 102 is composed of a plurality of hidden layers, and a plurality of neurons are arranged in each hidden layer.
- the output layer 103 is combined with the neuron information transmitted by the intermediate layer 102, and is output as the crushing strength of the painted steel pipe that is painted after the final steel pipe molding. Learning is performed by gradually optimizing the weighting coefficient in the neural network model based on the output result and the actual crushing strength of the read past coated steel pipe.
- the result storage unit 64 stores the training data, the parameters (weighting factors) of the neural network model, and the output result of the neural network model for the training data in the storage device 9.
- the steel pipe manufacturing characteristic calculation unit 7 of the calculation processing unit 5 uses the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6 to determine the steel pipe shape and steel pipe shape of the coated steel pipe to be predicted for the crush strength. By inputting the steel pipe strength characteristics after forming, the pipe making strain during steel pipe forming, and the steel pipe manufacturing characteristics consisting of the coating conditions, the crushing strength of the coated steel pipe coated after forming the steel pipe corresponding to the said steel pipe manufacturing characteristics is predicted. Perform processing. Then, when the steel pipe manufacturing characteristic determination mode information is the steel pipe manufacturing characteristic determination mode, the steel pipe manufacturing characteristic calculation unit 7 gradually approaches the predicted crushing strength of the coated steel pipe to the required target crushing strength of the coated steel pipe.
- At least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions, which form the steel pipe manufacturing characteristics, is sequentially changed to obtain the optimum steel pipe manufacturing characteristics. Performs the process of determining.
- the steel pipe manufacturing characteristic calculation unit 7 includes 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 as functional blocks. ing.
- the information reading unit 71 performs a process of reading the steel pipe crushing strength prediction model generated by the steel pipe crushing strength prediction model generation unit 6. Further, the information reading unit 71 is used to predict the crushing strength input to the steel pipe crushing strength prediction model, such as the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the steel pipe forming strain.
- Information on steel pipe manufacturing characteristics consisting of coating conditions, information on crushing strength of target coated steel pipes, and information on steel pipe manufacturing characteristic determination mode are read. Further, the crushing strength prediction unit 72 is based on the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions, which are the targets for predicting the crushing strength read by the information reading unit 71.
- the steel pipe manufacturing characteristics are input to the steel pipe crushing strength prediction model read by the information reading unit 71, and a process of predicting the crushing strength of the coated steel pipe painted after the steel pipe is formed is performed.
- the steel pipe manufacturing characteristic determination unit 73 and the crushing strength prediction unit 72 when the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is the steel pipe manufacturing characteristic determination mode, the predicted crushing strength of the coated steel pipe is determined.
- the steel pipe shape after steel pipe forming the steel pipe strength characteristic after steel pipe forming, the pipe making strain during steel pipe forming, and the coating conditions so as to gradually approach the required target crushing strength of the coated steel pipe. At least one is sequentially changed to determine the optimum steel pipe manufacturing characteristics, and a process of outputting the information of the determined optimum steel pipe manufacturing characteristics to the result output unit 74 is performed.
- the steel pipe manufacturing characteristic determination unit 73 information on the crushing strength of the coated steel pipe predicted by the crushing strength prediction unit 72 when the steel pipe manufacturing characteristic determination mode information read by the information reading unit 71 is not the steel pipe manufacturing characteristic determination mode.
- the process of outputting the predicted value) to the result output unit 74 is performed.
- the result output unit 74 performs a process of outputting the determined optimum steel pipe manufacturing characteristic information or the predicted crushing strength information (predicted value) of the coated steel pipe to the output device 10, and stores these information.
- the process of storing in the device 9 is performed.
- the steel pipe manufacturing characteristic calculation unit 7 inputs a steel pipe manufacturing characteristic calculation command to the input device 8 and receives the steel pipe manufacturing characteristic calculation command
- the steel pipe manufacturing characteristic calculation unit 7 executes the steel pipe manufacturing characteristic calculation program 42 stored in the ROM 4 and performs information.
- Each function of the reading unit 71, the crush strength prediction unit 72, the steel pipe manufacturing characteristic determination unit 73, and the result output unit 74 is executed.
- the information reading unit 71 of the steel pipe manufacturing characteristic calculation unit 7 reads the steel pipe crush strength prediction model generated by the steel pipe crush strength prediction model generation unit 6 stored in the storage device 9 in step S11.
- the information reading unit 71 reads information on the crushing strength of the coated steel pipe, which is input from a higher-level computer (not shown) and stored in the storage device 9 and is painted after forming the required target steel pipe. ..
- the information reading unit 71 is input to the input device 8 by the operator, and is input to the steel pipe crushing strength prediction model stored in the storage device 9. Information on steel pipe manufacturing characteristics including the later steel pipe shape, steel pipe strength characteristics after steel pipe forming, pipe forming strain during steel pipe forming, and coating conditions is read.
- step S14 the information reading unit 71 inputs to the input device 8 by the operator and stores the steel pipe manufacturing characteristic determination mode information (information on whether or not the mode determines the optimum steel pipe manufacturing characteristic) stored in the storage device 9. ) Is read.
- step S15 the crushing strength prediction unit 72 applies the crushing strength prediction model read in step S11 to the steel pipe shape and steel pipe after forming the crushing strength of the coated steel pipe to be predicted in step S13.
- the crushing strength of the coated steel pipe is predicted by inputting the steel pipe strength characteristics after forming, the pipe forming strain during steel pipe forming, and the steel pipe manufacturing characteristics consisting of the coating conditions.
- the steel pipe shape after forming the steel pipe of the coated steel pipe to be predicted is applied to the steel pipe crush strength prediction model generated by the method for generating the steel pipe crush strength prediction model according to the second embodiment of the present invention.
- the steel pipe manufacturing characteristic determination mode information (information on whether or not the mode is the mode for determining the optimum steel pipe manufacturing characteristic) read in step S14 is the steel pipe manufacturing characteristic determination mode (optimal). It is determined whether or not the mode is used to determine the manufacturing characteristics of steel pipes.
- step S16 When the determination result in step S16 is YES (in the steel pipe manufacturing characteristic determination mode), the process proceeds to step S17, and when the determination result in step S16 is NO (not in the steel pipe manufacturing characteristic determination mode), step S19. Move to.
- step S17 the steel pipe manufacturing characteristic determination unit 73 determines the crushing strength (predicted value) of the coated steel pipe predicted in step S15 and the crushing strength (target value) of the required target coated steel pipe read in step S12. It is determined whether or not the difference from the above is within a predetermined threshold.
- this predetermined threshold value is set to approximately 0.5% to 1%.
- step S17 When 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 predetermined threshold value), the process proceeds to step S18, and when the determination result in step S17 is NO (predicted value). When it is determined that the difference between the target value and the target value is larger than the predetermined threshold value), the process proceeds to step S20.
- step S20 the steel pipe manufacturing characteristic determination unit 73 determines the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the steel pipe forming time in the steel pipe manufacturing characteristics of the coated steel pipe to be predicted in the crushing strength read in step S13. At least one of the pipe forming strain and the coating condition is changed, and the process returns to step S15.
- the crushing strength prediction unit 72 has at least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions in step S20.
- the changed steel pipe manufacturing characteristics of the steel pipe are input to the steel pipe crushing strength prediction model read in step S11, and the crushing strength of the coated steel pipe is predicted again.
- the steel pipe manufacturing characteristic determination unit 73 determines the crushing strength (predicted value) of the coated steel pipe repredicted in step S15 and the required target read in step S12 in step S17. It is determined whether or not the difference from the crushing strength (target value) of the coated steel pipe is within a predetermined threshold. Then, a series of steps of step S20, step S15, step S16, and step S17 are repeatedly executed until the determination result becomes YES.
- step S17 when 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 predetermined threshold value), the process proceeds to step S18.
- the steel pipe manufacturing characteristic determination unit 73 determines the steel pipe shape after steel pipe forming when the difference between the predicted value and the target value is within a predetermined threshold value, the steel pipe strength characteristic after steel pipe forming, and the manufacturing during steel pipe forming.
- the steel pipe manufacturing characteristics consisting of pipe strain and coating conditions are determined as the optimum steel pipe manufacturing characteristics.
- step S16, step S17, step S20, step S15, step S16, step S17 and step S18 the crushing strength of the coated steel pipe painted after the predicted steel pipe forming according to the second embodiment of the present invention is determined.
- the steel pipe strength characteristic after steel pipe forming, the pipe making strain during steel pipe forming, and the coating conditions, which are included in the steel pipe manufacturing characteristics so as to gradually approach the required target crushing strength of the coated steel pipe At least one of the above is sequentially changed to correspond to a steel pipe manufacturing characteristic determination method for determining the optimum steel pipe manufacturing characteristics.
- step S19 the result output unit 74 of the steel pipe manufacturing characteristic calculation unit 7 determines the optimum steel pipe manufacturing characteristic determined in step S18 when the determination result in step S16 is YES (in the steel pipe manufacturing characteristic determination mode). Information is output to the output device 10.
- the result output unit 74 provides information on the crushing strength of the coated steel pipe to be painted after the steel pipe molding predicted in step S15 (prediction). Value) is output to the output device 10.
- the method for generating the steel pipe crushing strength prediction model according to the second embodiment of the present invention is based on the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions.
- the crushing strength of the steel pipe after forming the steel pipe is obtained by machine-learning a plurality of learning data using the past steel pipe manufacturing characteristics as input data and the crushing strength of the steel pipe after forming the past steel pipe as output data for this input data.
- Generate a predicted steel pipe crush strength prediction model (steel pipe crush strength prediction model generation unit 6).
- the method for predicting the crushing strength of a steel pipe according to the second embodiment of the present invention is based on the steel pipe crushing strength prediction model generated by the method for generating the steel pipe crushing strength prediction model, and the steel pipe after forming the steel pipe of the coated steel pipe to be predicted.
- the crushing strength of the steel pipe after steel pipe forming is predicted by inputting the shape, the steel pipe strength characteristic after steel pipe forming, the pipe making strain during steel pipe forming, and the steel pipe manufacturing characteristic including the coating conditions (steps S11 to S15).
- the crushing strength of the coated steel pipe that is painted after the steel pipe is formed can be predicted with high accuracy in consideration of the pipe forming strain during the steel pipe forming.
- the coating conditions that greatly affect the crushing strength of the coated steel pipe are taken into consideration when predicting the crushing strength of the coated steel pipe, the prediction accuracy of the crushing strength of the coated steel pipe can be further improved.
- the method for determining the manufacturing characteristics of the steel pipe according to the second embodiment of the present invention is included in the steel pipe manufacturing characteristics so that the predicted crushing strength of the coated steel pipe approaches the required crushing strength of the coated steel pipe.
- At least one of the steel pipe shape after steel pipe forming, the steel pipe strength characteristics after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions are sequentially changed to determine the optimum steel pipe manufacturing characteristics (step S16, Step S17, step S20, step S15, step S16, step S17 and step S18).
- 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 linked to the coated steel pipe formed in the forming step. That is, the method for manufacturing a steel pipe according to the second embodiment of the present invention includes a coated steel pipe forming step of forming a steel pipe and painting the formed steel pipe to form a coated steel pipe, and a method of predicting the crushing strength of the steel pipe (step S11). In steps S15), the crushing strength prediction step for predicting the crushing strength of the coated steel pipe formed in the coated steel pipe forming step and the crushing strength of the coated steel pipe predicted by the crushing strength prediction step were formed in the coated steel pipe forming step. It may be provided with a performance prediction value assigning step associated with the coated steel pipe.
- the predicted crushing strength of the coated steel pipe in the performance prediction value assigning step is associated with the coated steel pipe, for example, the predicted crushing strength (predicted value) of the coated steel pipe is imparted to the coated steel pipe by marking. This is achieved by attaching a tag stating the predicted crushing strength (predicted value) of the painted steel pipe to the painted steel pipe. As a result, the person who handles the coated steel pipe can grasp the crushing strength (predicted value) of the coated steel pipe.
- the manufacturing conditions of the painted steel pipe selection of the pipe making method, bending rate at the time of pipe making
- the painted steel pipe may be manufactured under the determined manufacturing conditions of the painted steel pipe.
- the method for manufacturing the steel pipe according to the second embodiment of the present invention was determined by the method for determining the manufacturing characteristics of the coated steel pipe (step S16, step S17, step S20, step S15, step S16, step S17 and step S18).
- the manufacturing conditions of the coated steel pipe may be determined based on the optimum steel pipe manufacturing characteristics, and the coated steel pipe may be manufactured under the determined manufacturing conditions of the coated steel pipe.
- the manufactured coated steel pipe satisfies the determined optimum steel pipe manufacturing characteristics, and as a result, the predicted crushing strength (predicted value) of the coated steel pipe is gradually approaching the required crushing strength of the coated steel pipe. It becomes a coated steel pipe with excellent pressure-resistant crushing performance, and it is possible to avoid damage or damage to the structure.
- the past steel pipe manufacturing characteristics that are input data when generating the steel pipe crushing strength prediction model are the steel pipe shape after steel pipe forming and the steel pipe forming after steel pipe forming. It is the steel pipe strength characteristics and the pipe making strain during steel pipe forming.
- the past steel pipe manufacturing characteristics may include the steel pipe shape after the past steel pipe forming, the steel pipe strength characteristics after the steel pipe forming, and the pipe forming strain during the steel pipe forming, and other past steel pipe manufacturing characteristics, for example, It may include the strength characteristics of the steel sheet before forming the steel pipe in the past.
- the past steel pipe manufacturing characteristics that are input data when generating the steel pipe crushing strength prediction model are the steel pipe shape after the past steel pipe forming and the steel pipe. It suffices to include the steel pipe strength characteristics after forming, the pipe forming strain during steel pipe forming, and the coating conditions, and may include other past steel pipe manufacturing characteristics, for example, the strength characteristics of the steel pipe before forming the past steel pipes. ..
- the steel pipe shape after the past steel pipe forming, which is the input data is the steel pipe.
- the steel pipe strength characteristic after forming the steel pipe, which is the input data is the steel pipe.
- Young's modulus E GPa
- Poisson's ratio ⁇ -
- ⁇ tensile strength YS
- MPa tensile strength
- YS tensile strength
- YS compression strength 0.23% YS of steel pipe
- the intensity is not limited to 0.5% YS (stress corresponding to 0.5% strain).
- the steel pipe crushing strength prediction model includes the steel pipe shape of the steel pipe to be predicted after the steel pipe is formed, the steel pipe strength characteristics after the steel pipe is formed, and the pipe making during the steel pipe forming.
- the steel pipe manufacturing characteristics consisting of strain are input.
- This steel pipe manufacturing characteristic may include the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, and the pipe forming strain during steel pipe forming, and other steel pipe manufacturing characteristics, for example, the strength of the steel sheet before steel pipe forming. You may enter the characteristic.
- the steel pipe crushing strength prediction model includes the steel pipe shape of the coated steel pipe to be predicted after the steel pipe is formed, the steel pipe strength characteristics after the steel pipe is formed, and the structure at the time of steel pipe forming.
- the steel pipe manufacturing characteristics consisting of pipe strain and coating conditions are input.
- This steel pipe manufacturing characteristic may include the steel pipe shape after steel pipe forming, the steel pipe strength characteristic after steel pipe forming, the pipe forming strain during steel pipe forming, and the coating conditions, and other steel pipe manufacturing characteristics, for example, before steel pipe forming.
- the strength characteristics of the steel plate may be input.
- the shape of the steel pipe after forming the steel pipe input to the steel pipe crushing strength prediction model is the maximum outer diameter Dmax (mm) of the steel pipe and the steel pipe. It is not limited to the minimum outer diameter Dmin (mm), the average outer diameter Dave (mm) of the steel pipe, the average plate thickness t (mm) of the steel pipe, and the roundness (Ovality) fO (%) of the outer diameter shape of the steel pipe.
- the steel pipe strength characteristics after forming the steel pipe which are input to the steel pipe crushing strength prediction model, are the Young's modulus E (GPa) of the steel pipe and Poisson's ratio of the steel pipe. Ratio ⁇ (-), tensile strength of steel pipe YS (MPa), compression strength of steel pipe 0.23% YS (stress corresponding to 0.23% strain), and compression strength of steel pipe 0.5% YS (0.5) It is not limited to the stress corresponding to% strain).
- the machine learning method is a neural network
- the steel tube crush strength prediction model is a prediction model constructed by a neural network, but any machine learning method may be used. For example, it may be a decision tree.
- the steel pipe crushing strength prediction model is used for the steel pipe shape after past steel pipe forming (maximum outer diameter Dmax (mm) of steel pipe, minimum outer diameter Dmin (mm) of steel pipe, average outer diameter Dave of steel pipe (Dave).
- the generated steel pipe crush strength prediction model shows the steel pipe shape (maximum outer diameter Dmax (mm) of the steel pipe, minimum outer diameter of the steel pipe) of the steel pipe to be predicted as shown in Table 1 after forming the steel pipe.
- Example 1 and 2 the actual crushing strength after steel pipe forming was actually measured from the steel pipe shape after steel pipe forming shown in Table 1, the steel pipe strength characteristics after steel pipe forming, and the pipe forming strain during steel pipe forming (actual). Tube test results).
- the criteria for determining the actual tube test results in Examples 1 and 2, Examples 3 to 6 and Comparative Examples 1 to 6 shown below are the same, and the difference between the actual crushing strength obtained in the experiment and the standard reference value is determined. Evaluated, NG if the actual crushing strength is below the standard standard value, C if the actual crushing strength is above the standard standard value within the range of less than 10%, and the actual crushing strength is higher than the standard standard value.
- B was defined as 10% or more and less than 20%
- A was defined as the actual crushing strength exceeding the standard value by 20% or more.
- the actual crushing strength (actual pipe test result) after forming the actually measured steel pipe was lower than the standard standard value (predetermined standard value), and the judgment result was NG, but the steel pipe crushing strength was predicted.
- the predicted value of the crushing strength after forming the steel pipe using the model was also lower than the standard standard value (predetermined standard value), and the judgment result was NG, and the experimental evaluation and the result were in agreement.
- the steel pipe crushing strength prediction model is used for the steel pipe shape after past steel pipe forming (maximum outer diameter Dmax (mm) of steel pipe, minimum outer diameter Dmin (mm) of steel pipe, average outer diameter of steel pipe).
- Dave (mm) average plate thickness t (mm) of steel pipe, roundness (Ovality) fO (%) of outer diameter shape of steel pipe, steel pipe strength characteristics after past steel pipe forming (Young rate E of steel pipe) GPa), Poisson ratio ⁇ (-) of steel pipe, tensile strength YS (MPa) of steel pipe, compression strength 0.23% YS (stress corresponding to 0.23% strain) of steel pipe, and compression strength 0.5 of steel pipe.
- the generated steel pipe crush strength prediction model is used to indicate the shape of the steel pipe after forming the steel pipe to be predicted as shown in Table 1 (maximum outer diameter Dmax (mm) of the steel pipe, minimum outside of the steel pipe).
- Diameter Dmin (mm) average outer diameter of steel pipe Dave (mm), average plate thickness of steel pipe t (mm), roundness (Ovality) fO (%) of outer diameter shape of steel pipe, steel pipe after steel pipe forming Strength characteristics (Young ratio E (GPa) of steel pipe, Poisson ratio ⁇ (-) of steel pipe, tensile strength YS (MPa) of steel pipe, compressive strength 0.23% YS of steel pipe (stress corresponding to 0.23% strain) , And the compression strength of the steel pipe 0.5% YS (stress corresponding to 0.5% strain)), the pipe making strain during steel pipe forming (tensile strain (%) during steel pipe forming), and the coating conditions (maximum temperature (maximum temperature ()).
- Example 3 the actual crushing strength (actual pipe test result) of the actually measured coated steel pipe exceeded the standard standard value (predetermined standard value) by 20% or more, and the determination result was A, and the steel pipe crushing strength.
- the predicted value of the crushing strength of the coated steel pipe using the prediction model also exceeded the standard standard value (predetermined standard value) by 20% or more, and the judgment result was A, which was in agreement with the experimental evaluation.
- Example 6 the actual crushing strength (actual pipe test result) of the actually measured coated steel pipe exceeds the standard standard value (predetermined standard value) in the range of 10% or more and less than 20%, and the judgment result is B.
- the predicted value of the crushing strength of the coated steel pipe using the steel pipe crushing strength prediction model also exceeds the standard standard value (predetermined standard value) within the range of 10% or more and less than 20%. bottom.
- the steel pipe shape (maximum outer diameter Dmax (mm) of the steel pipe, the steel pipe of the steel pipe to be predicted shown in Table 1 after forming the steel pipe).
- After forming the steel pipe by inputting the later steel pipe strength characteristics Young ratio E (GPa) of the steel pipe, Poisson ratio ⁇ (-) of the steel pipe, and tensile strength of the steel pipe (yield stress corresponding to 0.5% strain)). The crushing strength of was predicted.
- Minimum outer diameter Dmin (mm), average outer diameter Dave (mm) of steel pipe, average plate thickness t (mm) of steel pipe, roundness (Ovality) fO (%) of outer diameter shape of steel pipe, and steel pipe Enter the steel pipe strength characteristics after forming (Young ratio E (GPa) of the steel pipe, Poisson ratio ⁇ (-) of the steel pipe, and tensile strength of the steel pipe (yield stress corresponding to 0.5% strain)), and the coated steel pipe. The crushing strength of was predicted.
- Comparative Example 5 the actual crushing strength (actual pipe test result) of the actually measured coated steel pipe exceeded the predetermined standard value by 20% or more, and the determination result was A, but the prediction formula of Non-Patent Document 1 was used.
- the predicted value of the crushing strength of the coated steel pipe exceeded the standard standard value (predetermined standard value) in the range of 10% or more and less than 20%, and the judgment result was B.
- the result was inconsistent.
- Comparative Example 6 the actual crushing strength (actual pipe test result) of the actually measured coated steel pipe exceeded the standard standard value (predetermined standard value) in the range of less than 10%, and the judgment result was C, but it was not patented.
- the predicted value of the crushing strength of the coated steel pipe using the prediction formula of Document 1 exceeds the standard standard value (predetermined standard value) in the range of 10% or more and less than 20%, and the judgment result is B, and there is a difference between the two. It occurred, and the predicted value and the result of the experimental evaluation were inconsistent.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Mechanical Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Biochemistry (AREA)
- Automation & Control Theory (AREA)
- Computer Hardware Design (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Manufacturing & Machinery (AREA)
- Quality & Reliability (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
図1には、本発明の第1実施形態に係る鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、及び鋼管の製造特性決定方法が適用される鋼管製造特性決定装置の概略構成の機能ブロック図が示されている。
図1に示す第1実施形態に適用される鋼管製造特性決定装置1は、鋼管圧潰強度予測モデルの生成、及び生成された鋼管圧潰強度予測モデルを用いた鋼管成形後の鋼管の圧潰強度を予測を行う。また、鋼管製造特性決定装置1は、予測された鋼管成形後の鋼管の圧潰強度が要求される目標の鋼管成形後の鋼管の圧潰強度に漸近するような最適な鋼管製造特性の決定を行う。
ここで、鋼管は、一般的に、板状の鋼板を円管形状に曲げ加工して成形し製造され、その後、表面に塗装する場合もある。
なお、鋼管の圧潰強度とは、鋼管が圧潰するときの負荷応力(MPa)を意味し、ここでいう「圧潰」とは、負荷応力が最大値を示しこれ以上に外圧に対し形状を保てなくなるまで変形した状態をいうものとする。
ここで、鋼管圧潰強度予測モデル生成部6による機能の実現に必要な情報としては、例えば、鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみからなる過去の鋼管製造特性を入力データとし、この入力データに対する過去の鋼管成形後の鋼管の圧潰強度を出力データとした複数の学習用データが挙げられる。
演算処理部5は、機能ブロックとして、鋼管圧潰強度予測モデル生成部6と、鋼管製造特性演算部7とを備えている。
また、前処理部62は、学習用データ取得部61が取得した複数の学習用データを、鋼管圧潰強度予測モデル生成用に加工する。具体的には、前処理部62は、学習用データを構成する鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみからなる過去の鋼管製造特性の実績情報を、ニューラルネットワークモデルに読み込ませるために、0~1の間で標準化(正規化)を行う。
ニューラルネットワークモデルである鋼管圧潰強度予測モデルは、入力側から順に入力層101、中間層102、及び出力層103を含んでいる。
モデル生成部63が、ハイパーパラメータを用いたニューラルネットワークモデルによる学習を行うに際し、入力層101には、前処理部62で加工された学習用データを構成する鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみからなる過去の鋼管製造特性の実績情報、即ち、0~1の間で標準化された過去の鋼管製造特性の実績情報が格納される。
出力層103は、中間層102により伝達されたニューロンの情報が結合され、最終的な鋼管成形後の鋼管の圧潰強度として出力される。この出力した結果と、読み込まれた過去の鋼管成形後の鋼管の圧潰強度の実績とに基づき、ニューラルネットワークモデル内の重み係数が徐々に最適化されることで学習が行われる。
結果保存部64は、学習用データ、ニューラルネットワークモデルのパラメータ(重み係数)、及び学習用データに対するニューラルネットワークモデルの出力結果を、記憶装置9に記憶させる。
情報読取部71は、記憶装置9に記憶された鋼管製造特性演算部7による機能の実現に必要な情報を読み込む処理を行う。具体的に、情報読取部71は、鋼管圧潰強度予測モデル生成部6によって生成された鋼管圧潰強度予測モデルを読み込む処理を行う。また、情報読取部71は、鋼管圧潰強度予測モデルに入力される圧潰強度の予測対象となる鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみからなる鋼管製造特性の情報、目標の鋼管成形後の鋼管の圧潰強度の情報、及び鋼管製造特性決定モード情報を読み込む処理を行う。
また、結果出力部74は、決定された最適な鋼管製造特性の情報あるいは予測された鋼管成形後の鋼管の圧潰強度の情報(予測値)を出力装置10に出力する処理を行うとともに、これらの情報を記憶装置9に記憶させる処理を行う。
この鋼管製造特性演算部7は、入力装置8に鋼管製造特性の演算指令が入力され、鋼管製造特性の演算指令を受けると、ROM4に記憶されている鋼管製造特性演算プログラム42を実行し、情報読取部71、圧潰強度予測部72、鋼管製造特性決定部73、及び結果出力部74の各機能を実行する。
次いで、情報読取部71は、ステップS2において、図示しない上位計算機から入力され、記憶装置9に記憶されている、要求される目標の鋼管成形後の鋼管の圧潰強度の情報を読み込む。
次いで、情報読取部71は、ステップS4において、オペレータによって入力装置8に入力され、記憶装置9に記憶されている鋼管製造特性決定モード情報(最適な鋼管製造特性を決定するモードか否かの情報)を読み込む。
このステップS1~ステップS5は、本発明の第1実施形態に係る、鋼管圧潰強度予測モデルの生成方法により生成された鋼管圧潰強度予測モデルに、予測対象となる鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみからなる鋼管製造特性を入力して鋼管成形後の鋼管の圧潰強度を予測する鋼管の圧潰強度予測方法に対応する。
そして、ステップS6における判定結果がYESのとき(鋼管製造特性決定モードのとき)は、ステップS7に移行し、ステップS6における判定結果がNOのとき(鋼管製造特性決定モードでないとき)は、ステップS9に移行する。
ステップS7では、鋼管製造特性決定部73は、ステップS5で予測された鋼管成形後の鋼管の圧潰強度(予測値)と、ステップS2で読み込まれた、要求される目標の鋼管成形後の鋼管の圧潰強度(目標値)との差異が所定の閾値以内か否かを判定する。
ここで、前述の所定の閾値は、目標値や製造条件によって異なるがおおむね0.5%~1%に設定される。
ステップS10では、鋼管製造特性決定部73は、ステップS3で読み込まれた圧潰強度の予測対象となる鋼管の鋼管製造特性における鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみのうちの少なくとも1つを変更し、ステップS5に戻す。
このステップS6、ステップS7、ステップS10、ステップS5、ステップS6、ステップS7及びステップS8は、本発明の第1実施形態に係る、予測された鋼管成形後の鋼管の圧潰強度が、要求される目標の鋼管成形後の鋼管の圧潰強度に漸近するように、鋼管製造特性に含まれる鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみのうちの少なくとも一つを逐次変更し、最適な鋼管製造特性を決定する鋼管の製造特性決定方法に対応する。
これにより、鋼管製造特性演算部7の処理が終了する。
これにより、鋼管成形時の造管ひずみを考慮して鋼管成形後の鋼管の圧潰強度を精度高く予測するに際しての、鋼管圧潰強度予測モデルを適切に生成することができる。
これにより、鋼管成形時の造管ひずみを考慮して鋼管成形後の鋼管の圧潰強度を精度高く予測することができる。
これにより、予測された鋼管成形後の鋼管の圧潰強度が、要求される目標の鋼管成形後の鋼管の圧潰強度に漸近するときの、鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみからなる最適な鋼管製造特性を決定することができる。
つまり、本発明の第1実施形態に係る鋼管の製造方法は、鋼管を成形する鋼管の成形工程と、鋼管の圧潰強度予測方法(ステップS1~ステップS5)により、成形工程で成形された鋼管の圧潰強度を予測する圧潰強度予測工程と、圧潰強度予測工程により予測された鋼管の圧潰強度を成形工程で成形された鋼管に紐付ける性能予測値付与工程とを備えていてもよい。
これにより、成形された鋼管を取り扱う者は、当該鋼管の圧潰強度(予測値)を把握することができる。
つまり、本発明の第1実施形態に係る鋼管の製造方法は、鋼管の製造特性決定方法(ステップS6、ステップS7、ステップS10、ステップS5、ステップS6、ステップS7及びステップS8)により決定された最適な鋼管製造特性に基づいて鋼管の製造条件を決定し、その決定された鋼管の製造条件で鋼管を製造するようにしてもよい。
これにより、製造された鋼管が、決定された最適な鋼管製造特性を満足し、その結果、予測される鋼管の圧潰強度(予測値)が要求される目標の鋼管成形後の鋼管の圧潰強度に漸近したものとなり、耐圧潰性能に優れた鋼管となり、構造物の損傷や損壊事故を回避することができる。
本発明の第2実施形態に係る鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法について、図1及び図4乃至図5を参照して説明する。第1実施形態において既に説明した部材については説明を省略することがある。
図1に示す鋼管製造特性決定装置1は、第2実施形態に係る鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、及び鋼管の製造特性決定方法にも適用される。第2実施形態に係る鋼管圧潰強度予測モデルの生成方法は、鋼管成形後に塗装してなる塗装鋼管の鋼管圧潰強度予測モデルの生成を行う。第2実施形態に係る鋼管の圧潰強度予測方法は、生成された鋼管圧潰強度予測モデルを用いて鋼管成形後に塗装してなる塗装鋼管の圧潰強度を予測する。第2実施形態に係る鋼管の製造特性決定方法は、予測された塗装鋼管の圧潰強度が要求される目標の塗装鋼管の圧潰強度に漸近するような最適な鋼管製造特性の決定を行う。
入力装置8には、第1実施形態と同様に、鋼管圧潰強度予測モデルの生成指令、鋼管製造特性の演算指令等が入力される。第2実施形態では、第1実施形態と異なり、鋼管成形後に塗装してなる塗装鋼管の圧潰強度を予測するので、鋼管製造特性として、圧潰強度の予測対象となる塗装鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみの他に塗装条件が入力される。また、入力装置8には、目標の鋼管形後に塗装してなる塗装鋼管の圧潰強度等が入力される。
成形した鋼管に塗装を施すのは、防食のためであり、特に海底パイプラインで使用する鋼管は耐食性を優れたものとするため、成形後に塗装を施すのが一般的である。この塗装における塗装条件(最高温度(℃)及び保持時間(min))は、鋼管成形後の鋼管強度特性に影響し、塗装鋼管の圧潰性能に直接影響するため、入力装置8に入力するようにした。塗装のコーティング加熱の影響により、鋼管の材質が変化(転位の堆積・回復・ひずみ時効など)することで、鋼管成形後の鋼管の圧潰強度(塗装前の圧潰性能)からその圧潰強度が増加又は低下する。
また、出力装置10は、演算装置2からの出力データ、例えば、圧潰強度予測部72で予測された鋼管成形後に塗装してなる塗装鋼管の圧潰強度(予測値)の情報や鋼管製造特性決定部73で決定された最適な鋼管製造特性の情報を出力する出力ポートとして機能する。
演算処理部5の鋼管圧潰強度予測モデル生成部6は、記憶装置9に記憶された、鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる過去の鋼管製造特性を入力データとし、この入力データに対する過去の鋼管成形後に塗装してなる塗装鋼管の圧潰強度を出力データとした複数の学習用データを機械学習させて、鋼管圧潰強度予測モデルを生成する。機械学習の手法は、第1実施形態と同様に、ニューラルネットワークであり、鋼管圧潰強度予測モデルは、ニューラルネットワークにより構築された予測モデルである。
また、前処理部62は、第1実施形態と同様に、学習用データ取得部61が取得した複数の学習用データを、鋼管圧潰強度予測モデル生成用に加工する。
ニューラルネットワークモデルである鋼管圧潰強度予測モデルは、入力側から順に入力層101、中間層102、及び出力層103を含んでいる。
モデル生成部63が、ハイパーパラメータを用いたニューラルネットワークモデルによる学習を行うに際し、入力層101には、前処理部62で加工された学習用データを構成する鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる過去の鋼管製造特性の実績情報、即ち、0~1の間で標準化された過去の鋼管製造特性の実績情報が格納される。
出力層103は、中間層102により伝達されたニューロンの情報が結合され、最終的な鋼管成形後に塗装してなる塗装鋼管の圧潰強度として出力される。この出力した結果と、読み込まれた過去の塗装鋼管の圧潰強度の実績とに基づき、ニューラルネットワークモデル内の重み係数が徐々に最適化されることで学習が行われる。
結果保存部64は、学習用データ、ニューラルネットワークモデルのパラメータ(重み係数)、及び学習用データに対するニューラルネットワークモデルの出力結果を、記憶装置9に記憶させる。
情報読取部71は、鋼管圧潰強度予測モデル生成部6によって生成された鋼管圧潰強度予測モデルを読み込む処理を行う。また、情報読取部71は、鋼管圧潰強度予測モデルに入力される圧潰強度の予測対象となる鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる鋼管製造特性の情報、目標の塗装鋼管の圧潰強度の情報、及び鋼管製造特性決定モード情報を読み込む処理を行う。
また、圧潰強度予測部72は、情報読取部71で読み込んだ圧潰強度の予測対象となる鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる鋼管製造特性を、情報読取部71で読み込んだ鋼管圧潰強度予測モデルに入力して、鋼管成形後に塗装してなる塗装鋼管の圧潰強度を予測する処理を行う。
また、結果出力部74は、決定された最適な鋼管製造特性の情報あるいは予測された塗装鋼管の圧潰強度の情報(予測値)を出力装置10に出力する処理を行うとともに、これらの情報を記憶装置9に記憶させる処理を行う。
この鋼管製造特性演算部7は、入力装置8に鋼管製造特性の演算指令が入力され、鋼管製造特性の演算指令を受けると、ROM4に記憶されている鋼管製造特性演算プログラム42を実行し、情報読取部71、圧潰強度予測部72、鋼管製造特性決定部73、及び結果出力部74の各機能を実行する。
次いで、情報読取部71は、ステップS12において、図示しない上位計算機から入力され、記憶装置9に記憶されている、要求される目標の鋼管成形後に塗装してなる塗装鋼管の圧潰強度の情報を読み込む。
次いで、情報読取部71は、ステップS13において、オペレータによって入力装置8に入力され、記憶装置9に記憶されている鋼管圧潰強度予測モデルに入力される圧潰強度の予測対象となる塗装鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる鋼管製造特性の情報を読み込む。
その後、圧潰強度予測部72は、ステップS15において、ステップS11で読み込まれた鋼管圧潰強度予測モデルに、ステップS13で読み込まれた圧潰強度の予測対象となる塗装鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる鋼管製造特性を入力して、塗装鋼管の圧潰強度を予測する。
続いて、鋼管製造特性決定部73は、ステップS16において、ステップS14で読み込んだ鋼管製造特性決定モード情報(最適な鋼管製造特性を決定するモードか否かの情報)が鋼管製造特性決定モード(最適な鋼管製造特性を決定するモード)か否かを判定する。
ステップS17では、鋼管製造特性決定部73は、ステップS15で予測された塗装鋼管の圧潰強度(予測値)と、ステップS12で読み込まれた、要求される目標の塗装鋼管の圧潰強度(目標値)との差異が所定の閾値以内か否かを判定する。
ここで、この所定の閾値は、おおむね0.5%~1%に設定される。
ステップS20では、鋼管製造特性決定部73は、ステップS13で読み込まれた圧潰強度の予測対象となる塗装鋼管の鋼管製造特性における鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件のうちの少なくとも1つを変更し、ステップS15に戻す。
これにより、鋼管製造特性演算部7の処理が終了する。
これにより、鋼管成形時の造管ひずみを考慮して鋼管成形後に塗装してなる塗装鋼管の圧潰強度を精度高く予測するに際しての、鋼管圧潰強度予測モデルを適切に生成することができる。
また、塗装鋼管の圧潰強度を予測する鋼管圧潰強度予測モデルの生成に際し、塗装鋼管の圧潰強度に大きな影響を与える塗装条件も考慮しているから、鋼管圧潰強度予測モデルの精度もより高くすることができる。
これにより、鋼管成形時の造管ひずみを考慮して鋼管成形後に塗装してなる塗装鋼管の圧潰強度を精度高く予測することができる。
そして、塗装鋼管の圧潰強度を予測するに際して塗装鋼管の圧潰強度に大きな影響を与える塗装条件も考慮しているから、塗装鋼管の圧潰強度の予測精度をより高めることができる。
これにより、予測された塗装鋼管の圧潰強度が、要求される目標の塗装鋼管の圧潰強度に漸近するときの、鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件からなる最適な鋼管製造特性を決定することができる。
つまり、本発明の第2実施形態に係る鋼管の製造方法は、鋼管を成形し、成形された鋼管に塗装して塗装鋼管を形成する塗装鋼管形成工程と、鋼管の圧潰強度予測方法(ステップS11~ステップS15)により、塗装鋼管形成工程で形成された塗装鋼管の圧潰強度を予測する圧潰強度予測工程と、圧潰強度予測工程により予測された塗装鋼管の圧潰強度を塗装鋼管形成工程で形成された塗装鋼管に紐付ける性能予測値付与工程とを備えていてもよい。
これにより、塗装鋼管を取り扱う者は、当該塗装鋼管の圧潰強度(予測値)を把握することができる。
また、塗装鋼管を製造するに際して、出力装置10で出力されたステップS18で決定された最適な鋼管製造特性の情報に基づいて塗装鋼管の製造条件(造管方法の選択、造管時の曲げ率、造管時のひずみ量、塗装時の昇温速度、塗装時の最高到達温度、塗装時の最高到達温度保持時間、塗装時の最高到達温度保持時間経過後の冷却速度など)を決定し、その決定された塗装鋼管の製造条件で塗装鋼管を製造するようにしてもよい。
これにより、製造された塗装鋼管が、決定された最適な鋼管製造特性を満足し、その結果、予測される塗装鋼管の圧潰強度(予測値)が要求される目標の塗装鋼管の圧潰強度に漸近したものとなり、耐圧潰性能に優れた塗装鋼管となり、構造物の損傷や損壊事故を回避することができる。
例えば、第1実施形態に係る鋼管圧潰強度予測モデルの生成方法において、鋼管圧潰強度予測モデルを生成する際に、入力データとなる過去の鋼管製造特性は、鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみとしている。しかし、過去の鋼管製造特性は、過去の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみを含めばよく、それら以外の過去の鋼管製造特性、例えば、過去の鋼管成形前の鋼板の強度特性を含んでいても良い。
また、第1実施形態及び第2実施形態に係る鋼管圧潰強度予測モデルの生成方法おいて、鋼管圧潰強度予測モデルを生成する際に、入力データとなる鋼管成形後の鋼管強度特性は、鋼管のヤング率E(GPa)、鋼管のポアソン比μ(-)、鋼管の引張強度YS(MPa)、鋼管の圧縮強度0.23%YS(0.23%歪に対応する応力)、及び鋼管の圧縮強度0.5%YS(0.5%歪に対応する応力)に限られない。
また、第1実施形態及び第2実施形態に鋼管の圧潰強度予測方法において、鋼管圧潰強度予測モデルに入力される鋼管成形後の鋼管強度特性は、鋼管のヤング率E(GPa)、鋼管のポアソン比μ(-)、鋼管の引張強度YS(MPa)、鋼管の圧縮強度0.23%YS(0.23%歪に対応する応力)、及び鋼管の圧縮強度0.5%YS(0.5%歪に対応する応力)に限られない。
また、第1実施形態及び第2実施形態において、機械学習の手法はニューラルネットワークであり、鋼管圧潰強度予測モデルは、ニューラルネットワークにより構築された予測モデルとしてあるが、機械学習法であればよく、例えば決定木などであってもよい。
また、実施例3~6では、表1に示す鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件から塗装鋼管の圧潰強度を実測した(実管試験結果)。
また、比較例2では、実測した鋼管成形後の実際の圧潰強度(実管試験結果)が規格基準値(所定規格値)を下回り判定結果はNGであったが、非特許文献1の予測式を用いた鋼管成形後の圧潰強度の予測値は規格基準値(所定規格値)を10%未満の範囲で上回り判定結果はCであり、両者の間に差異が生じ、予測値と実験評価との結果が不一致であった。
また、比較例4では、実測した塗装鋼管の実際の圧潰強度(実管試験結果)が規格基準値(所定規格値)を20%以上上回り判定結果はAであったが、非特許文献1の予測式を用いた塗装鋼管の圧潰強度の予測値は規格基準値(所定規格値)を10%以上かつ20%未満の範囲で上回り判定結果はBであり、両者の間に差異が生じ、予測値と実験評価との結果が不一致であった。
また、比較例6では、実測した塗装鋼管の実際の圧潰強度(実管試験結果)が規格基準値(所定規格値)を10%未満の範囲で上回り判定結果はCであったが、非特許文献1の予測式を用いた塗装鋼管の圧潰強度の予測値は規格基準値(所定規格値)を10%以上かつ20%未満の範囲で上回り判定結果はBであり、両者の間に差異が生じ、予測値と実験評価との結果が不一致であった。
2 演算装置
3 RAM
4 ROM
5 演算処理部
6 鋼管圧潰強度予測モデル生成部
7 鋼管製造特性演算部
8 入力装置
9 記憶装置
10 出力装置
11 バス
41 鋼管圧潰強度予測モデル生成プログラム
42 鋼管製造特性演算プログラム
61 学習用データ取得部
62 前処理部
63 モデル生成部
64 結果保存部
71 情報読取部
72 圧潰強度予測部
73 鋼管製造特性決定部
74 結果出力部
101 入力層
102 中間層
103 出力層
Claims (12)
- 鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみを含む過去の鋼管製造特性を入力データとし、この入力データに対する過去の鋼管成形後の鋼管の圧潰強度を出力データとした複数の学習用データを機械学習させて、鋼管成形後の鋼管の圧潰強度を予測する鋼管圧潰強度予測モデルを生成することを特徴とする鋼管圧潰強度予測モデルの生成方法。
- 前記機械学習の手法はニューラルネットワークであり、前記鋼管圧潰強度予測モデルは、ニューラルネットワークにより構築された予測モデルであることを特徴とする請求項1に記載の鋼管圧潰強度予測モデルの生成方法。
- 請求項1又は2に記載の鋼管圧潰強度予測モデルの生成方法により生成された鋼管圧潰強度予測モデルに、予測対象となる鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみを含む鋼管製造特性を入力して鋼管成形後の鋼管の圧潰強度を予測することを特徴とする鋼管の圧潰強度予測方法。
- 請求項3に記載された鋼管の圧潰強度予測方法により予測された鋼管成形後の鋼管の圧潰強度が、要求される目標の鋼管成形後の鋼管の圧潰強度に漸近するように、鋼管製造特性に含まれる鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、及び鋼管成形時の造管ひずみのうちの少なくとも一つを逐次変更し、最適な鋼管製造特性を決定することを特徴とする鋼管の製造特性決定方法。
- 鋼管を成形する鋼管の成形工程と、請求項3に記載の鋼管の圧潰強度予測方法により、前記成形工程で成形された鋼管の圧潰強度を予測する圧潰強度予測工程と、該圧潰強度予測工程により予測された鋼管の圧潰強度を前記成形工程で成形された鋼管に紐付ける性能予測値付与工程とを備えることを特徴とする鋼管の製造方法。
- 請求項4に記載の鋼管の製造特性決定方法により決定された最適な鋼管製造特性に基づいて鋼管の製造条件を決定し、その決定された鋼管の製造条件で鋼管を製造することを特徴とする鋼管の製造方法。
- 鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件を含む過去の鋼管製造特性を入力データとし、この入力データに対する過去の鋼管成形後に塗装してなる塗装鋼管の圧潰強度を出力データとした複数の学習用データを機械学習させて、鋼管成形後に塗装してなる塗装鋼管の圧潰強度を予測する鋼管圧潰強度予測モデルを生成することを特徴とする鋼管圧潰強度予測モデルの生成方法。
- 前記機械学習の手法はニューラルネットワークであり、前記鋼管圧潰強度予測モデルは、ニューラルネットワークにより構築された予測モデルであることを特徴とする請求項7に記載の鋼管圧潰強度予測モデルの生成方法。
- 請求項7又は8に記載の鋼管圧潰強度予測モデルの生成方法により生成された鋼管圧潰強度予測モデルに、予測対象となる塗装鋼管の鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件を含む鋼管製造特性を入力して鋼管成形後に塗装してなる塗装鋼管の圧潰強度を予測することを特徴とする鋼管の圧潰強度予測方法。
- 請求項9に記載された鋼管の圧潰強度予測方法により予測された塗装鋼管の圧潰強度が、要求される目標の塗装鋼管の圧潰強度に漸近するように、鋼管製造特性に含まれる鋼管成形後の鋼管形状、鋼管成形後の鋼管強度特性、鋼管成形時の造管ひずみ、及び塗装条件のうちの少なくとも一つを逐次変更し、最適な鋼管製造特性を決定することを特徴とする鋼管の製造特性決定方法。
- 鋼管を成形し、成形された鋼管に塗装して塗装鋼管を形成する塗装鋼管形成工程と、請求項9に記載の鋼管の圧潰強度予測方法により、前記塗装鋼管形成工程で形成された塗装鋼管の圧潰強度を予測する圧潰強度予測工程と、該圧潰強度予測工程により予測された塗装鋼管の圧潰強度を前記塗装鋼管形成工程で形成された塗装鋼管に紐付ける性能予測値付与工程とを備えることを特徴とする鋼管の製造方法。
- 請求項10に記載の鋼管の製造特性決定方法により決定された最適な鋼管製造特性に基づいて塗装鋼管の製造条件を決定し、その決定された塗装鋼管の製造条件で塗装鋼管を製造することを特徴とする鋼管の製造方法。
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US17/925,762 US20230191466A1 (en) | 2020-05-26 | 2021-02-08 | Steel pipe collapse strength prediction model generation method, steel pipe collapse strength prediction method, steel pipe manufacturing characteristics determination method, and steel pipe manufacturing method |
KR1020227040334A KR20230002856A (ko) | 2020-05-26 | 2021-02-08 | 강관 압궤 강도 예측 모델의 생성 방법, 강관의 압궤 강도 예측 방법, 강관의 제조 특성 결정 방법 및, 강관의 제조 방법 |
EP21813944.2A EP4160182A4 (en) | 2020-05-26 | 2021-02-08 | METHOD FOR GENERATION OF MODEL FOR PREDICTING CRUSHING RESISTANCE OF STEEL PIPE, METHOD FOR PREDICTING CRUSHING RESISTANCE OF STEEL PIPE, METHOD FOR DETERMINING MANUFACTURING CHARACTERISTICS OF STEEL PIPE, AND METHOD FOR MANUFACTURING PIPE IN STEEL |
BR112022023698A BR112022023698A2 (pt) | 2020-05-26 | 2021-02-08 | Método de geração de modelo de previsão de resistência de colapso de tubo de aço, método de previsão de resistência de colapso de tubo de aço, método de determinação de características de fabricação de tubo de aço e método de fabricação de tubo de aço |
CN202180037995.7A CN115667875A (zh) | 2020-05-26 | 2021-02-08 | 钢管压溃强度预测模型的生成方法、钢管的压溃强度预测方法、钢管的制造特性决定方法以及钢管的制造方法 |
JP2021519178A JP7103514B2 (ja) | 2020-05-26 | 2021-02-08 | 鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2020-091127 | 2020-05-26 | ||
JP2020091127 | 2020-05-26 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021240900A1 true WO2021240900A1 (ja) | 2021-12-02 |
Family
ID=78744252
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/004588 WO2021240900A1 (ja) | 2020-05-26 | 2021-02-08 | 鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 |
Country Status (7)
Country | Link |
---|---|
US (1) | US20230191466A1 (ja) |
EP (1) | EP4160182A4 (ja) |
JP (1) | JP7103514B2 (ja) |
KR (1) | KR20230002856A (ja) |
CN (1) | CN115667875A (ja) |
BR (1) | BR112022023698A2 (ja) |
WO (1) | WO2021240900A1 (ja) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN118190633B (zh) * | 2024-05-17 | 2024-07-16 | 云南通衢工程检测有限公司 | 一种钢绞线钢丝应力测试仪应力测试控制***及方法 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001349883A (ja) * | 2000-06-09 | 2001-12-21 | Hitachi Metals Ltd | 金属材料の特性予測方法 |
WO2003099482A1 (fr) * | 2002-05-24 | 2003-12-04 | Nippon Steel Corporation | Tuyau en acier uoe presentant une excellente resistance aux impacts, et procede de fabrication du tuyau en acier uoe |
JP2014222160A (ja) * | 2013-05-13 | 2014-11-27 | Jfeスチール株式会社 | 鋼板の曲げ加工後における、加工方向と直交する方向の引張特性の推定方法 |
WO2015030210A1 (ja) * | 2013-08-30 | 2015-03-05 | 新日鐵住金株式会社 | 耐サワー性、耐圧潰特性及び低温靭性に優れた厚肉高強度ラインパイプ用鋼板とラインパイプ |
WO2018074433A1 (ja) * | 2016-10-18 | 2018-04-26 | 新日鐵住金株式会社 | 圧潰強度予測方法 |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH06259107A (ja) * | 1993-03-02 | 1994-09-16 | Kobe Steel Ltd | プロセスラインにおける学習制御方法 |
US7132617B2 (en) | 2002-02-20 | 2006-11-07 | Daimlerchrysler Corporation | Method and system for assessing quality of spot welds |
CN101701315B (zh) | 2009-10-30 | 2011-06-08 | 中海石油金洲管道有限公司 | 海底管线钢管的制造方法 |
-
2021
- 2021-02-08 CN CN202180037995.7A patent/CN115667875A/zh active Pending
- 2021-02-08 US US17/925,762 patent/US20230191466A1/en active Pending
- 2021-02-08 EP EP21813944.2A patent/EP4160182A4/en active Pending
- 2021-02-08 KR KR1020227040334A patent/KR20230002856A/ko unknown
- 2021-02-08 WO PCT/JP2021/004588 patent/WO2021240900A1/ja active Application Filing
- 2021-02-08 BR BR112022023698A patent/BR112022023698A2/pt unknown
- 2021-02-08 JP JP2021519178A patent/JP7103514B2/ja active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001349883A (ja) * | 2000-06-09 | 2001-12-21 | Hitachi Metals Ltd | 金属材料の特性予測方法 |
WO2003099482A1 (fr) * | 2002-05-24 | 2003-12-04 | Nippon Steel Corporation | Tuyau en acier uoe presentant une excellente resistance aux impacts, et procede de fabrication du tuyau en acier uoe |
JP2014222160A (ja) * | 2013-05-13 | 2014-11-27 | Jfeスチール株式会社 | 鋼板の曲げ加工後における、加工方向と直交する方向の引張特性の推定方法 |
WO2015030210A1 (ja) * | 2013-08-30 | 2015-03-05 | 新日鐵住金株式会社 | 耐サワー性、耐圧潰特性及び低温靭性に優れた厚肉高強度ラインパイプ用鋼板とラインパイプ |
WO2018074433A1 (ja) * | 2016-10-18 | 2018-04-26 | 新日鐵住金株式会社 | 圧潰強度予測方法 |
Non-Patent Citations (4)
Title |
---|
DJERRAD ABDERRAHIM; FAN FENG; ZHI XU-DONG; WU QI-JIAN: "Artificial Neural Networks (ANN) Based Compressive Strength Prediction of AFRP Strengthened Steel Tube", INTERNATIONAL JOURNAL OF STEEL STRUCTURES, vol. 20, no. 1, 9 September 2019 (2019-09-09), pages 156 - 174, XP037006504 * |
FUJII, KATASHI ET AL.: "A Prediction Method of Strength Deterioration in Aging of Circular Steel Tube Corroded in Marine Environment", JOURNAL OF JAPAN SOCIETY OF CIVIL ENGINEERS, vol. 66, no. 1, March 2010 (2010-03-01), pages 92 - 105, XP055877965 * |
OFFSHORE STANDARD DNV-OS-F101, SUBMARINE PIPELINE SYSTEMS, DET NORSKE VERITAS, vol. 5, October 2010 (2010-10-01), pages 41 - 56 |
See also references of EP4160182A4 |
Also Published As
Publication number | Publication date |
---|---|
CN115667875A (zh) | 2023-01-31 |
EP4160182A4 (en) | 2023-11-22 |
EP4160182A1 (en) | 2023-04-05 |
KR20230002856A (ko) | 2023-01-05 |
BR112022023698A2 (pt) | 2022-12-20 |
US20230191466A1 (en) | 2023-06-22 |
JPWO2021240900A1 (ja) | 2021-12-02 |
JP7103514B2 (ja) | 2022-07-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Sathish | GAC-ANN technique for prediction of spring back effect in wipe bending process of sheet metal | |
Liu et al. | Deep learning in sheet metal bending with a novel theory-guided deep neural network | |
CN108563906B (zh) | 一种基于深度学习的短纤维增强复合材料宏观性能预测法 | |
Inamdar et al. | Studies on the prediction of springback in air vee bending of metallic sheets using an artificial neural network | |
CN106096073A (zh) | 一种基于损伤力学非概率区间分析模型的金属疲劳裂纹全寿命预估方法 | |
JP6969713B1 (ja) | 鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 | |
Mirzaali et al. | Optimization of tube hydroforming process using simulated annealing algorithm | |
Rossi et al. | Testing methodologies for the calibration of advanced plasticity models for sheet metals: A review | |
CN107180123B (zh) | 一种高强度钢潜水器耐压球壳极限承载力估算方法 | |
WO2021240900A1 (ja) | 鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 | |
Şenol et al. | Springback analysis in air bending process through experiment based artificial neural networks | |
Fan et al. | Prediction algorithm for springback of frame-rib parts in rubber forming process by incorporating Sobol within improved grey relation analysis | |
Bosetti et al. | Identification of johnson–cook and tresca's parameters for numerical modeling of aisi-304 machining processes | |
Ma et al. | Machine learning (ML)-based prediction and compensation of springback for tube bending | |
Ingarao et al. | Optimization methods for complex sheet metal stamping computer aided engineering | |
Xu et al. | Prediction of springback in local bending of hull plates using an optimized backpropagation neural network | |
Zhang et al. | Identification of constitutive parameters for thin-walled aluminium tubes using a hybrid strategy | |
WO2022054336A1 (ja) | 鋼管圧潰強度予測モデルの生成方法、鋼管の圧潰強度予測方法、鋼管の製造特性決定方法、及び鋼管の製造方法 | |
Chen | An analysis of forming limit in the elliptic hole-flanging process of sheet metal | |
Lin et al. | Optimization of bending process parameters for seamless tubes using Taguchi method and finite element method | |
Strano et al. | Hierarchical metamodeling of the air bending process | |
Pham et al. | Process parameter optimization for incremental forming of aluminum alloy 5052-H32 sheets using back-propagation neural network | |
Baig et al. | Machine Learning for the Prediction of Springback in High Tensile Strength Steels after V-Bending Process Using Tree-Based Learning | |
Barati et al. | The effect of normal anisotropy on thin-walled tube bending | |
Lee et al. | A study on fatigue damage modeling using neural networks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2021519178 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21813944 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 20227040334 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202217066230 Country of ref document: IN |
|
REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112022023698 Country of ref document: BR |
|
ENP | Entry into the national phase |
Ref document number: 112022023698 Country of ref document: BR Kind code of ref document: A2 Effective date: 20221121 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2021813944 Country of ref document: EP Effective date: 20230102 |