Manufacturing method of thick steel plate for welded structure, and method for generating weld defect occurrence prediction model

Abstract

To provide a thick steel sheet for welded structure and method for manufacturing the same, a method for creating a weld defect generation prediction model, and a welded structure, which allow suppression of low temperature cracks by controlling segregation of a sheet thickness center part.SOLUTION: Provided is a thick steel sheet for a welded structure, having a component composition containing C, Si, Mn, P, S, Al, Ti and N and consisting of Fe and inevitable impurities, and in the thick steel sheet, the area ratio of a Mn-component concentrated part in a sheet thickness center part is 3% or less; the Charpy absorption energy vE0 at a test temperature of 0°C in a sheet thickness direction at a sheet thickness center position is 27 J or more; the yield strength is 325 to 750 N / mm2; and the yield strength ratio is 90% or less with a tensile strength of 490 to 930 N / mm2. A method for manufacturing the thick steel sheet for the welded structure has the following steps: on the basis of a weld defect generation prediction model, inputting welding specifications and acquiring manufacturing specifications for the thick steel sheet such that a defect generation rate in a welded part gradually approaches a desired value or is within a range of the desired value; and manufacturing the thick steel sheet by using the manufacturing specifications for the thick steel sheet.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a thick steel plate used in welded structures in the fields of civil engineering, architecture, and bridges, which has excellent weldability over a wide range of conditions, with a heat input of 15 to 1200 kJ / cm, a yield strength of 325 MPa or more, a tensile strength of 490 MPa or more, a yield ratio of 90% or less, and an average Charpy absorbed energy of 27 J or more at a test temperature of 0°C in the thickness direction at the center of the plate thickness; a method for manufacturing such a thick steel plate; a method for generating a model for predicting the occurrence of weld defects; and a welded structure using such a thick steel plate. [Background technology]

[0002] In recent years, as welded structures have become larger, steel plates have become stronger and thicker. At the same time, there is a demand for improved welding efficiency to improve the construction efficiency of structures and reduce construction costs, and the scope of application of high-heat input welding is expanding. For example, ultra-high heat input welding methods such as submerged arc welding and electroslag welding, with welding heat inputs exceeding 400 kJ / cm, are used for box columns used in high-rise buildings. For example, columns used in high-rise building structures are constructed by welding four long steel plates (skin plates) together at the corners of their long sides to form a box-shaped column. Additionally, diaphragms are welded inside the columns where beams are to be welded to ensure strength. Single-pass welding is desirable for these welding processes to improve construction efficiency and reduce construction time. Therefore, high-heat-input submerged arc welding (HSAW) with a heat input of up to 600 kJ / cm is used for corner welding between skin plates. High-heat-input electroslag welding with a heat input of 400–1200 kJ / cm is increasingly used for welding skin plates and diaphragms. However, high-heat-input welding can lead to coarsening of the metal structure in the heat-affected zone (HAZ), which can reduce the toughness of the weld and can also increase the risk of cold cracking and delayed fracture. In particular, corner welding, which is the welding between skin plates, is perpendicular to the direction of extension of the center segregation in the steel plate for the skin plate. In other words, force is applied in the direction most susceptible to cracking relative to the center segregation, increasing the possibility of cold cracking.

[0003] To solve these problems, various steel plates have been proposed to improve HAZ toughness.

[0004] For example, Patent Document 1 discloses a technique in which 0.20 to 0.60% Mo is added to steel containing 0.07 to 0.09% C, which makes it easy to ensure the strength of the base metal and HAZ, to make the HAZ structure of high heat input welding a bainite single phase, and further reduce Si and P to improve HAZ toughness. This technique is said to enable the production of high-strength thick steel plates with excellent toughness in the heat-affected zone (HAZ) and resistance to weld cracking.

[0005] Patent Document 2 discloses that nitrides such as TiN and BN are finely dispersed by controlling the alloying elements to Ti: 0.003-0.02%, B: 0.0005-0.0030%, Ca: 0.0015-0.0030%, and N: 0.0040-0.008%. The document introduces a technology that suppresses coarse Ca-containing inclusions and elongated MnS-based inclusions that are detrimental to through-thickness toughness, and further suppresses central segregation of C, thereby ensuring through-thickness (Z-direction) toughness in the weld heat-affected zone (HAZ). [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-208213 [Patent Document 2] Japanese Patent Application Laid-Open No. 2009-221522 Summary of the Invention [Problem to be solved by the invention]

[0007] However, the prior art disclosed in the above patent document has the following problems.

[0008] Patent Document 1 does not mention the toughness in the thickness direction, which is the current issue.

[0009] Patent Document 2 describes that since it is difficult to determine the toughness in the thickness direction, the quality is ensured by manufacturing control without actually evaluating it. Since neither of these technologies directly evaluates the toughness in the thickness direction, an accurate solution to the cold cracking problem could not be found.

[0010] The present invention has been developed in consideration of the above-mentioned problems of the conventional technology, and aims to provide a thick steel plate for welded structures that can suppress cold cracking by controlling segregation in the center of the plate thickness, a manufacturing method thereof, a method for generating a weld defect occurrence prediction model, and a welded structure. [Means for solving the problem]

[0011] In order to solve the above problems, the inventors have conducted extensive research into the requirements for steel plate center segregation and cold cracking in the weld heat affected zone (HAZ) in thick steel plates for welded structures. As a result, it was discovered that reducing the area of ​​center segregation in steel plates is the key to producing welded structures without generating welding defects, including cold cracking. However, it was also confirmed that the threshold for whether or not a weld defect occurs and the manufacturing specifications of various steel plates that affect the area of ​​center segregation are not determined by a single factor, but are influenced by complex factors. Therefore, by constructing a model using machine learning and statistical analysis with large amounts of data, it was possible to take into account complex influences, and we discovered a manufacturing method for welded structures that does not produce welding defects, which is our ultimate goal.

[0012] The thick steel plate for welded structures according to the present invention, which was developed based on the above findings, has the following configuration. [1] A steel sheet containing, by mass%, C: 0.03 to 0.16%, Si: 0.50% or less, Mn: 0.8 to 3.0%, P: 0.015% or less, S: 0.0050% or less, Al: 0.005 to 0.100%, Ti: 0.004 to 0.030%, N: 0.0015 to 0.0065%, with the balance being Fe and unavoidable impurities, the area ratio of the Mn-enriched portion at the center of the thickness is 3% or less, the Charpy absorbed energy vE0 at the center of the thickness in the thickness direction at a test temperature of 0°C is 27J or more, and the yield strength is 325 to 750N / mm 2 , and tensile strength is 490 to 930 N / mm 2 The steel plate for welded structures has a yield ratio of 90% or less. Here, the Mn-enriched portion is defined as a region in an analytical field including the center of the plate thickness where the Mn concentration in the steel plate is 1.2 times or more the Mn element analysis value of the molten steel in the ladle. [2] In the above [1], the steel plate for welded structures further contains, as the chemical composition, one or more selected from, by mass%, Cu: 0.01 to 1.00%, Ni: 0.01 to 2.50%, Cr: 1.5% or less, Mo: 1.0% or less, Nb: 0.1% or less, V: 0.2% or less, Ca: 0.005% or less, REM: 0.02% or less, Mg: 0.005% or less, and B: 0.005% or less.

[0013] The method for manufacturing a thick steel plate for a welded structure according to the present invention, which was developed based on the above findings, and the method for generating a prediction model for weld defect occurrence during the manufacturing of the thick steel plate, are configured as follows. [3] A method for manufacturing a thick steel plate for a welded structure according to [1] or [2] above, comprising: a step of acquiring manufacturing specifications for a thick steel plate, using training data including the results of the manufacturing specifications of the thick steel plate, the mechanical property values ​​of the thick steel plate, the results of the welding specifications, and the results of the occurrence of defects in the welds, based on a weld defect occurrence prediction model that has been trained in advance by machine learning, to input specific welding specifications and acquire manufacturing specifications for the thick steel plate so that the defect occurrence rate in the welds approaches a desired value or is within the desired value range; and a manufacturing step of manufacturing the thick steel plate using the manufacturing specifications for the thick steel plate acquired in the manufacturing specifications for the thick steel plate acquisition step. [4] A method for generating a prediction model for weld defect occurrence during the manufacturing of thick steel plates for welded structures as described in [1] or [2] above, which uses training data including the results of the manufacturing specifications of the thick steel plates, the mechanical property values ​​of the thick steel plates, the results of the welding specifications, and the results of the occurrence of defects in welds, to learn through machine learning and generate a prediction model for the occurrence of defects in welds.

[0014] The welded structure according to the present invention, which was developed based on the above findings, is configured as follows. [5] A welded structure that is single-layer or multi-layer arc-welded using the thick steel plate for welded structures described in [1] or [2] above. [6] In the above [5], the welded structure is a box column, and is a welded structure in which the thick steel plate is used as the steel plate on the flange side and is welded by single-layer or multi-layer submerged arc welding. [Effects of the Invention]

[0015] According to the present invention, it is possible to manufacture thick steel plates with a thickness range of up to 100 mm that have a yield strength of 325 MPa or more, a tensile strength of 490 MPa or more with a low yield ratio of 90% or less, and toughness with an average Charpy absorbed energy of 27 J or more at the center of the thickness direction at a test temperature of 0°C, and that have excellent weld cracking resistance in high heat input weld heat affected zones and low heat input multi-pass welds. This will greatly contribute to improving the efficiency of welding in welded structures and will have a significant effect on the industry. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a schematic diagram of a welded joint at a corner of a welded structure. DETAILED DESCRIPTION OF THE INVENTION

[0017] Hereinafter, a thick steel plate for a welded structure according to this embodiment will be described. <Chemical composition of steel plate> The steel plate has a chemical composition, in mass%, of C: 0.03 to 0.16%, Si: 0.50% or less, Mn: 0.8 to 3.0%, P: 0.015% or less, S: 0.0050% or less, Al: 0.005 to 0.100%, Ti: 0.004 to 0.030%, and N: 0.0015 to 0.0065%. Each component is explained below. In the following explanation, "%" representing the content of a component means "mass%."

[0018] C: 0.03% or more and 0.16% or less Carbon (C) is an element that increases the strength of steel and is useful for ensuring the strength required for structural steel. To minimize the amount of other alloying elements added, the C content is set to 0.03% or more. On the other hand, if the C content exceeds 0.16%, the weld crack resistance and HAZ toughness will decrease significantly. Therefore, the C content is set to a range of 0.03% to 0.16%.

[0019] Si:0.50% or less Si is an element that functions as a deoxidizer and also has the effect of increasing the strength of the base material. To achieve this effect, the Si content is preferably 0.01% or more. On the other hand, if the Si content exceeds 0.50%, the formation of island martensite is promoted, resulting in a significant decrease in toughness and weldability. Therefore, the Si content is set to 0.50% or less. Preferably, the Si content is 0.35% or less.

[0020] Mn: 0.8% or more and 3.0% or less Mn is an element that has the effect of increasing the strength of steel. The Mn content is set to 0.8% or more because it reduces island martensite in the microstructure of the heat-affected zone of high-heat-input welding, resulting in a finer structure, thereby ensuring toughness and ensuring a base metal yield strength of 325 MPa or more. Preferably, the Mn content is set to 1.5% or more. On the other hand, if the Mn content exceeds 3.0%, the toughness of the base metal and the toughness of the weld heat-affected zone will deteriorate significantly. Therefore, the Mn content is set to 3.0% or less. Preferably, the Mn content is set to 2.8% or less.

[0021] P:0.015% or less P concentrates in island martensite in the HAZ structure and promotes the formation of island martensite, thereby reducing HAZ toughness. Therefore, it is preferable to reduce P in order to improve HAZ toughness. Therefore, the P content is set to 0.015% or less. Preferably, the P content is set to 0.006% or less, which significantly improves HAZ toughness.

[0022] S: 0.0050% or less S is an element that deteriorates the low-temperature toughness of the base material, and it is preferable to reduce it as much as possible. If the S content exceeds 0.0050%, the deterioration of low-temperature toughness becomes significant, so the S content is set to 0.0050% or less. Preferably, the S content is 0.0030% or less.

[0023] Al: 0.005% or more and 0.100% or less Al is an element that acts as a deoxidizer and is most commonly used in the molten steel deoxidation process for high-tensile steel. Furthermore, Al fixes N in steel as AlN, contributing to improving the toughness of the base material. To achieve this effect, the Al content is set to 0.005% or more. Preferably, the Al content is set to 0.010% or more. On the other hand, if the Al content exceeds 0.100%, the toughness of the base material decreases, and Al is mixed into the weld metal during welding, deteriorating the toughness of the weld metal. Therefore, the Al content is set to 0.100% or less. Preferably, the Al content is set to 0.070% or less.

[0024] Ti: 0.004% or more and 0.030% or less Ti has a strong affinity with N and precipitates as TiN during solidification. The pinning effect of TiN, which is stable even at high temperatures, can suppress coarsening of austenite grains in the heat affected zone of high heat input welding, thereby improving the toughness of the weld heat affected zone. To achieve this effect, the Ti content is set to 0.004% or more. Preferably, the Ti content is set to 0.006% or more. On the other hand, if the Ti content exceeds 0.030%, TiN particles become coarse and the effect of suppressing coarsening of austenite grains becomes saturated. Therefore, the Ti content is set to 0.030% or less. Preferably, the Ti content is set to 0.025% or less.

[0025] N: 0.0015% or more and 0.0065% or less N is an element necessary to ensure TiN, and if it is less than 0.0015%, a sufficient amount of TiN cannot be ensured. Therefore, the N content is set to 0.0015% or more. Preferably, the N content is set to 0.0030% or more. On the other hand, if the N content exceeds 0.0065%, the amount of dissolved N increases, significantly reducing the toughness of the base metal and weld. Therefore, the N content is set to 0.0065% or less. Preferably, the N content is set to 0.0060% or less.

[0026] The above is the basic composition of the steel plate according to the embodiment, but optionally, one or more components selected from the following composition may further be contained. Cu: 0.01~1.00%, Ni: 0.01~2.50%, Cr: 1.5% or less, Mo: 1.0% or less, Nb: 0.1% or less, V: 0.2% or less, Ca: 0.005% or less, REM: 0.02% or less, Mg: 0.005% or less, B: 0.005% or less.

[0027] Cu: 0.01% or more and 1.00% or less Cu is an element that can increase strength while maintaining high toughness. In addition, Cu has little effect on the toughness of the heat-affected zone in high-heat-input welds, making it a useful element for increasing strength. When Cu is contained, the Cu content is set to 0.01% or more to obtain the above effects. Preferably, the Cu content is set to 0.10% or more, and more preferably, 0.20% or more. On the other hand, if the Cu content exceeds 1.00%, hot embrittlement occurs, deteriorating the surface properties of the steel sheet, so the Cu content is set to 1.00% or less. Preferably, the Cu content is set to 0.70% or less.

[0028] Ni: 0.01% or more and 2.50% or less Like Cu, Ni is an element that can increase strength while maintaining high toughness. In addition, Ni has little effect on the toughness of the heat-affected zone in high-heat-input welding, making it a useful element for increasing strength. When Ni is contained, the Ni content is set to 0.01% or more to obtain the above-mentioned effects. Preferably, the Ni content is set to 0.10% or more, and more preferably, 0.20% or more. On the other hand, if the Ni content exceeds 2.50%, the effect of addition saturates, and no effect commensurate with the content can be expected, which is economically disadvantageous. Therefore, the Ni content is set to 2.50% or less. Preferably, the Ni content is set to 1.7% or less.

[0029] Cr:1.5% or less Cr is an element that contributes to improving the strength of steel and can be added at any amount depending on the desired strength. However, if the Cr content exceeds 1.5%, the toughness of the heat-affected zone in high-heat-input welds deteriorates, so if Cr is added, the Cr content should be 1.5% or less. From the perspective of obtaining the strength-improving effect of Cr, the Cr content should preferably be 0.05% or more.

[0030] Mo: 1.0% or less Like Cr, Mo is an element that contributes to improving the strength of steel and can be added at any amount depending on the desired strength. However, if the Mo content exceeds 1.0%, the toughness of the heat-affected zone in high-heat-input welds deteriorates, so if Mo is added, the Mo content should be 1.0% or less. From the perspective of obtaining the strength-improving effect of Mo, the Mo content is preferably 0.05% or more.

[0031] Nb: 0.1% or less Like Cr and Mo, Nb is an element that contributes to improving the strength of steel and can be added at any amount depending on the desired strength. However, if the Nb content exceeds 0.1%, the toughness of the base material and the toughness of the heat-affected zone in high-heat-input welds deteriorates, so if Nb is added, the Nb content should be 0.1% or less. From the perspective of obtaining the strength-improving effect of Nb, the Nb content should preferably be 0.005% or more.

[0032] V: 0.2% or less Like Cr, Mo, and Nb, V is an element that contributes to improving the strength of steel and can be added at any amount depending on the desired strength. However, if the V content exceeds 0.2%, the toughness of the heat-affected zone in high-heat-input welds deteriorates, so if V is added, the V content should be 0.2% or less. From the perspective of obtaining the strength-improving effect of V, the V content is preferably 0.01% or more.

[0033] Ca: 0.005% or less Ca is an element that has the effect of improving toughness by refining crystal grains, and can be added at any amount depending on the desired properties. However, if the Ca content exceeds 0.005%, the effect of adding Ca becomes saturated, so if Ca is added, the Ca content should be 0.005% or less. From the viewpoint of obtaining the toughness-improving effect of Ca, the Ca content is preferably 0.001% or more.

[0034] REM: 0.02% or less Like Ca, REM (rare earth metals) have the effect of improving toughness, and can be added at any amount depending on the desired properties. However, if the REM content exceeds 0.02%, the effect of adding REM becomes saturated, so if REM is added, the REM content should be 0.02% or less. From the perspective of obtaining the effect of REM in improving toughness, the REM content is preferably 0.002% or more.

[0035] Mg: 0.005% or less Like Ca, Mg is an element that improves toughness by refining crystal grains, and can be added at any amount depending on the desired properties. However, if the Mg content exceeds 0.005%, the effect of adding Mg becomes saturated, so if Mg is added, the Mg content should be 0.005% or less. From the perspective of obtaining the toughness-improving effect of Mg, the Mg content is preferably 0.001% or more.

[0036] B: 0.005% or less B is an element that improves the hardenability and thereby improves the strength of steel. Furthermore, B has the effect of improving toughness during high heat input welding by fixing dissolved nitrogen as nitrides in the heat-affected zone of welds. However, if the B content exceeds 0.005%, the hardenability becomes excessively high, and the toughness and ductility of the base material decrease. Therefore, when B is contained, the B content is set to 0.005% or less. Preferably, the B content is set to 0.002% or less. From the viewpoint of obtaining the effect of adding B, the B content is preferably set to 0.0003% or more.

[0037] Ceq: 0.90% or less The carbon equivalent (Ceq) is calculated as C + Mn / 6 + Si / 24 + Ni / 40 + Cr / 5 + Mo / 4 + V / 14, and various alloying elements are preferably added according to the desired strength. If Ceq exceeds 0.90%, the weld crack resistance and HAZ toughness will be significantly reduced. Therefore, it is preferable to keep Ceq at 0.90% or less. Pcm: 0.35% or less The weld crack susceptibility composition Pcm is calculated as C + Si / 30 + Mn / 20 + Cu / 20 + Ni / 60 + Cr / 20 + Mo / 15 + V / 10 + 5 B. If Pcm exceeds 0.35%, the cold cracking susceptibility increases and cracks tend to occur in the weld metal, so it is preferable to keep Pcm at 0.35% or less. The chemical composition of the steel plate for welded structures according to this embodiment contains the above elements, with the balance being Fe and unavoidable impurities.

[0038] <Metal structure and mechanical properties of thick steel plates> Next, the metal structure of the steel plate for a welded structure according to this embodiment will be described. Area ratio of Mn-enriched area in the center of the plate thickness: 3% or less The metal structure of the steel sheet is not specified, but the area of ​​the element-enriched portion in the center of the sheet thickness is reduced. Here, the Mn-enriched portion is defined as a region in the analysis field including the center of the sheet thickness where the Mn concentration in the steel sheet is 1.2 times or more the Mn element analysis value of the molten steel in the ladle. Test pieces were taken from the width center and thickness center of the steel plate so that the cross sections were in the longitudinal direction and thickness direction, respectively, and the metal structure was evaluated. The chemical elements of the test piece were analyzed using an electron probe microanalyzer (EPMA). Chemical element analysis covered a 5mm x 10mm field of view in the thickness direction, centered at the center of the plate thickness. After electrolytic polishing of the surface of the test piece in the specified field of view, the Mn concentration was measured under the conditions of an acceleration voltage of 20kV, a 20μm long strip beam shape, and a 20μm step. Of the 250 x 500 points within the field of view, the positions where the Mn concentration was 1.2 times or more the Mn element analysis value of the molten steel in the ladle were defined as element-enriched areas (including inclusions), and the area ratio of the element-enriched areas was calculated for 10,250 points.

[0039] When an element-enriched area exists in the center of the plate thickness, the alloy element in the element-enriched area is high, resulting in high hardenability. As a result, the transformation temperature also decreases. In a thermal stress analysis using FEM (finite element method), it was confirmed that when the physical properties at the center of the plate thickness are changed and the temperature at which thermal expansion occurs due to transformation is lowered, thermal stress increases. Therefore, when corner welding is performed using a steel plate in which the ratio of the area of ​​the Mn-enriched portion in the center of the plate thickness exceeds 3%, there is a significantly high possibility of cold cracking occurring, so the ratio of the area of ​​the Mn-enriched portion is set to 3% or less.

[0040] Moreover, the thick steel plate for a welded structure according to this embodiment refers to a steel plate having a plate thickness of 12 mm or more.

[0041] Charpy absorbed energy vE0: 27J or more at a test temperature of 0°C in the thickness direction at the center of the plate The toughness of the steel plate for welded structures according to this embodiment is evaluated at the center of the plate thickness. The Charpy impact toughness vE0 (Charpy impact test value at 0°C) in the thickness direction when a V-notch is made at the center of the plate thickness is 27 J or more. The corners of the box column were welded, and the thermal stress in the weld was analyzed using the finite element method (FEM). As a result, it was found that the heat input from welding generates thermal stress, and that large thermal stress is applied in the thickness direction at the center of the plate thickness on the flange side. Therefore, if the toughness vE0 in the thickness direction at the center of the plate thickness is less than 27J, there is a high risk of cold cracking occurring at the center of the plate thickness on the flange side due to thermal stress caused by welding heat input. Therefore, the toughness vE0 in the thickness direction at the center of the plate thickness is 27J or more. Note that when the plate thickness is 55mm or less, it is difficult to take a Charpy test piece in the thickness direction. In that case, the notch position of the Charpy test piece is set at the center of the plate thickness, and separate steel material is pressed onto both ends to create a test piece, making it possible to perform the test.

[0042] Yield strength: 325N / mm 2 More than 750N / mm 2 below The yield strength (YS) of the steel plate for welded structures according to this embodiment is not particularly limited and can be set to any value. In consideration of use as a box column in a building structure, the yield strength of the steel plate is set to 325 N / mm 2 More than 750N / mm 2 The following applies.

[0043] Tensile strength: 490N / mm 2 More than 930N / mm 2 below The tensile strength (TS) of the steel plate for welded structures according to this embodiment is not particularly limited and can be set to any value. In consideration of use as a box column in a building structure, the tensile strength of the steel plate is set to 490 N / mm 2 More than 930N / mm 2 The following applies.

[0044] Yield ratio: 90% or less The yield ratio (YR) of the steel plate for welded structures according to this embodiment is not particularly limited and can be any value. Improved earthquake resistance is required for architectural structures, and the yield ratio (YR) is set to a low value of 90% or less to ensure the plastic deformability of the steel plate base material. When earthquake resistance is required, the yield ratio (YR) is preferably 85% or less, and more preferably 80% or less. Here, the yield ratio refers to the ratio of the yield strength (YS) to the tensile strength (TS) expressed as a percentage, that is, YR(%)=(YS / TS)×100.

[0045] <Method of manufacturing thick steel plates> Next, a method for manufacturing a thick steel plate for a welded structure according to this embodiment will be described.

[0046] Regarding the steel plate for welded structures having a low yield ratio and high tensile strength according to this embodiment, the method for producing the steel material (hot-rolled material) having the above-mentioned chemical composition is not particularly limited, and it can be produced, for example, by melting molten steel having the above-mentioned composition and casting it. The melting can be carried out by any method, such as a converter, an electric furnace, or an induction furnace. From the viewpoint of productivity, casting after melting is preferably carried out by a continuous casting method, but it can also be carried out by an ingot casting-breaking rolling method. For example, a steel slab can be used as the steel material. The thickness of the cast slab is preferably 500 mm or less. In order to reduce the proportion of the area of ​​the element-enriched portion at the center of the thickness of the steel plate, it is preferable that the casting conditions are strictly controlled, the columnar crystal structure is formed up to the center of the thickness of the slab, and soft reduction is applied. The columnar crystal structure can be observed by checking the macrostructure of the cross section of the slab.

[0047] The temperature of the steel material refers to the plate thickness average temperature. In production control, the plate thickness is sometimes divided into multiple sections and the temperatures of the steel plate surface, the 1 / 4t position, and the plate thickness center are controlled, but in the notation of this embodiment, when the plate thickness is divided into multiple sections and controlled, the plate thickness average temperature is the average of the temperatures of all the sections.

[0048] The thick steel plate for welded structures according to this embodiment is a thick steel plate with excellent weldability, and is manufactured by hot rolling without heat treatment. Note that, since the toughness in the thickness direction at the center of the plate thickness and the area ratio of the element-enriched portion at the center of the plate thickness do not change significantly due to heat treatment, heat treatment may be performed after hot rolling to obtain the desired yield strength, tensile strength, and yield ratio.

[0049] The above-mentioned steel material is heated prior to hot rolling. Heating may be carried out after the steel material obtained by a method such as casting has been cooled, or the steel material obtained by casting may be directly subjected to heating without being cooled. The microstructure and properties of the steel plate can be controlled by heat treatment after hot rolling. If the heating temperature of the steel material is less than 900°C, the deformation resistance of the steel material is high, which increases the load on the rolling mill during hot rolling and may make hot rolling difficult. Therefore, the heating temperature is preferably 900°C or higher. On the other hand, if the heating temperature is higher than 1250°C, oxidation of the steel becomes significant, increasing losses due to oxidation and resulting in a decrease in yield. Therefore, the heating temperature is preferably 1250°C or lower. After heating, the heated steel material is hot rolled to produce a thick steel plate. The final thickness of the thick steel plate is 19 mm or more and 100 mm or less.

[0050] After hot rolling, the steel plate may be cooled according to the yield strength, tensile strength, and yield ratio. The cooling conditions are not particularly limited, and cooling can be performed by any method, such as air cooling or water cooling. As water cooling, any cooling method using water (e.g., spray cooling, mist cooling, laminar cooling, etc.) can be used. The cooling temperature is not particularly limited, and can be, for example, from room temperature (e.g., 20°C) to 700°C.

[0051] Next, quality control in the method for manufacturing thick steel plates for welded structures according to this embodiment will be described. Although an example of the manufacturing method of the steel plate according to the present embodiment has been described above, the toughness in the thickness direction, which indicates the susceptibility to cold cracking during welding, may not be obtained even within the above-mentioned range of conditions. On the other hand, the toughness in the thickness direction changes due to the mutual influence of the manufacturing conditions in each process, so that it is very difficult to manage the operation so as to improve the toughness in the thickness direction.

[0052] Therefore, in order to reduce the proportion of the area of ​​the concentrated element portion at the center of the thickness of a steel plate, which affects the toughness in the plate thickness direction, it is effective to reduce the area of ​​the concentrated element portion at the center of the thickness of the plate at the stage of the slab after casting. The area of ​​the concentrated element portion is affected by the chemical composition, the width and thickness of the molten steel casting mold, the casting speed, the cooling conditions during casting, the soft reduction conditions, the electromagnetic stirring conditions, etc. Therefore, in order to reduce the area of ​​the concentrated element portion at the slab stage, it is effective to make the structure at the time of casting solidification columnar to the center of the slab thickness and to apply a soft reduction to the slab at the final stage of solidification. However, estimating the conditions under which columnar crystals are formed to the center of the slab thickness and predicting when the final stage of solidification occurs are difficult due to the many influencing factors.

[0053] Furthermore, when hot rolling a slab after casting, by increasing the ratio of the slab thickness to the steel plate thickness and performing hot rolling at a higher temperature, or by increasing the reduction per pass, the crystal grains in the center of the plate thickness become finer even if the area of ​​the element-enriched portion of the slab remains the same, thereby improving the toughness in the plate thickness direction.

[0054] As mentioned above, it is necessary to control the various manufacturing operation factors that interact in a complex manner. Therefore, a weld defect occurrence prediction model is utilized to manufacture steel plates that suppress cold cracking during welding in the construction of welded structures by controlling segregation in the center of the plate thickness. Therefore, a quality control method in a manufacturing method of thick steel plate will be described, but the present invention is not limited to this embodiment as long as the object is achieved.

[0055] This embodiment is a method for manufacturing thick steel plates, which includes a thick steel plate manufacturing specification acquisition step in which specific welding specifications are input based on a weld defect occurrence prediction model that has been trained in advance by machine learning using training data including the actual manufacturing specifications of the thick steel plates, mechanical property values ​​of the thick steel plates, actual welding specifications, and actual defect occurrence in welds, and the manufacturing specifications of the thick steel plates are acquired so that the defect occurrence rate in welds approaches a desired value or is within the range of the desired value, and a manufacturing step in which thick steel plates are manufactured using the manufacturing specifications of the thick steel plates acquired in the thick steel plate manufacturing specification acquisition step.

[0056] First, a weld defect occurrence prediction model trained by machine learning is constructed using training data including the actual manufacturing specifications of the thick steel plate, the mechanical property values ​​of the thick steel plate, the actual welding specifications, and the actual occurrence of defects in the weld.

[0057] Here, the actual manufacturing specifications of the steel plate used as training data include manufacturing performance data for the casting process and the rolling process. The manufacturing performance data for the casting process include the chemical composition, molten steel casting mold width, molten steel casting mold thickness, molten steel casting speed, cooling conditions during casting, soft reduction conditions, and electromagnetic stirring conditions, which affect the area of ​​the element-enriched portion at the center of the steel plate thickness. The manufacturing performance data for the rolling process include the heating temperature, controlled rolling start temperature, controlled rolling reduction ratio, finish rolling temperature, controlled cooling start temperature, controlled cooling end temperature, whether or not heat treatment was performed, the type of heat treatment, the heating temperature during heat treatment, and the temperature after heat treatment.

[0058] In addition, the mechanical property values ​​of the thick steel plate used as training data include the product plate thickness, plate thickness direction toughness value, yield strength, tensile strength, yield ratio, elongation, dimensions, shape, and area ratio of steel plate component enriched parts.

[0059] The welding specification results used as training data include the welding method, groove shape, root gap, preheat temperature, current, voltage, gas flow rate, welding speed, number of passes, and post-heat treatment conditions.

[0060] Furthermore, the actual defect occurrences in welds used as training data include cold cracking, hot cracking, and solidification cracking.

[0061] Here, as the prediction model of this embodiment, a statistical method and a machine learning model such as local regression, support vector machine, neural network, and random forest are created. Specifically, multiple or single models are used, and the most accurate combination or the most accurate model is selected.

[0062] Next, these actual results are linked as training data to construct a weld defect occurrence prediction model, and specific welding specifications are input to obtain an estimated weld defect occurrence rate that can be used to compare with a target weld defect occurrence rate. The welding specifications to be entered include the welding method, groove shape, root gap, type of welding material, preheat temperature, current, voltage, gas flow rate, welding speed, number of passes, and post-heat treatment conditions.

[0063] As the next step for obtaining manufacturing specifications for the thick steel plate, manufacturing specifications for the thick steel plate are searched for and obtained so that the calculated estimated value of the defect occurrence rate of the weld asymptotically approaches the desired value for the defect occurrence rate of the weld set as the target defect occurrence rate of the weld, or is within the desired value range. Here, the optimized mechanical properties of the steel plate can also be obtained.

[0064] Next, in the step of manufacturing the thick steel plate, the manufacturing specifications for the thick steel plate acquired in the step of acquiring the manufacturing specifications for the thick steel plate are used to set manufacturing conditions for each manufacturing line, and the thick steel plate is manufactured.

[0065] Through the above-described steel plate manufacturing steps, steel plate manufacturing specifications for manufacturing steel plates with optimized mechanical properties are obtained, and by using the steel plates manufactured through these manufacturing steps to manufacture welded structures, welded structures with a low defect rate in the welds can be obtained.

[0066] Another embodiment is a method for acquiring mechanical property values ​​of thick steel plates, which, when manufacturing thick steel plates for welded structures, uses training data including the mechanical property values ​​of the thick steel plates, the welding specification results, and the occurrence of defects in welds, and inputs specific welding specifications based on a weld defect occurrence prediction model that has been trained in advance by machine learning, to acquire mechanical property values ​​of the thick steel plates so that the defect occurrence rate in welds asymptotically approaches a desired value or is within the range of the desired value.

[0067] First, a weld defect occurrence prediction model trained by machine learning is generated using training data including the mechanical property values ​​of the steel plate, the welding specification results, and the weld defect occurrence results. Next, based on the weld defect occurrence prediction model linked to these results as training data, a specific welding specification is input and an estimated weld defect occurrence rate is obtained that is used to compare with a target weld defect occurrence rate.

[0068] Furthermore, the mechanical property values ​​of the thick steel plate are searched for and obtained so that the calculated estimated value of the defect occurrence rate of the weld asymptotically approaches the desired value for the defect occurrence rate of the weld set as the target defect occurrence rate of the weld, or is within the desired value range.

[0069] Here, the target defect occurrence rate of a weld is the target defect occurrence rate of a weld in a welded structure manufactured by welding a welding material having the mechanical properties of a specific thick steel plate according to specific welding specifications.

[0070] The mechanical property values ​​of thick steel plates used as training data to construct a weld defect occurrence prediction model include product plate thickness, toughness value in the plate thickness direction, yield strength, tensile strength, yield ratio, elongation, dimensions, shape, and area ratio of steel plate component enrichment areas.

[0071] The welding specification results used as training data include the welding method, groove shape, root gap, preheat temperature, current, voltage, gas flow rate, welding speed, number of passes, and post-heat treatment conditions.

[0072] Furthermore, the actual defect occurrences in welds used as training data include cold cracking, hot cracking, and solidification cracking.

[0073] It is preferable that the training data for constructing the weld defect occurrence prediction model further includes production performance data for the casting process and rolling process of the steel material in the production of thick steel plate.

[0074] The mechanical property values ​​of the steel plate to be predicted are mainly the toughness value in the thickness direction, but it is preferable to also predict one or more of the yield strength, tensile strength, elongation, yield ratio, longitudinal toughness value, dimensions, shape, defects, etc. The reason for this is that when optimizing the operating conditions, the optimized conditions that provide good toughness in the thickness direction may not satisfy other target properties, so it is preferable to construct a prediction model that also predicts mechanical property values ​​other than the toughness value in the thickness direction.

[0075] Here, as the prediction model of this embodiment, a statistical method and a machine learning model such as local regression, support vector machine, neural network, and random forest are created. Specifically, multiple or single models are used, and the most accurate combination or the most accurate model is selected.

[0076] The following describes a method for optimizing the mechanical property values ​​of thick steel plates, including the thickness direction toughness value obtained by the weld defect occurrence prediction model that also predicts mechanical property values ​​other than the above-mentioned thickness direction toughness value.

[0077] First, optimization of material properties is performed using the following formula 1. In formula 1, x is the design condition expressed as a vector, k is the type of property, fk(x) is the predicted value of the property, and αk is a preset weighting coefficient. The function fk(x) of the predicted value of the property in the evaluation function is based on the constructed prediction model. It is a set of design conditions x that satisfy the constraints. Therefore, the optimum design conditions are searched for within the range that satisfies the constraints. k are the lower and upper bounds of the characteristic value, respectively. Such optimization problems are solved using methods such as metaheuristics, genetic algorithms, mathematical programming, and swarm intelligence.

[0078]

number

[0079] If the optimal design conditions corresponding to the desired properties satisfy the constraints shown in Equation 1, the design conditions obtained by inverse analysis can be effectively used even if there are limits to the amount of additives added to the steel material and the capacity of the manufacturing equipment from the perspective of manufacturing costs. By setting constraints, it is possible to efficiently search within the range of the constraints rather than searching blindly. As described above, by manufacturing steel plates under the manufacturing conditions discovered above, it is possible to manufacture steel plates having, for example, a thickness direction toughness value of 27 J or more and target values ​​for mechanical properties other than the thickness direction toughness value.

[0080] <Welded structures> The steel plates for welded structures manufactured in the above-described embodiments can be used to construct welded structures by single-layer or multi-layer arc welding. In the case of welded box columns, the steel plates can be used as flange-side steel plates by single-layer or multi-layer submerged arc welding to construct the box columns. [Example]

[0081] The effects of this embodiment will be specifically described below based on examples, but the present invention is not limited to these examples. Steel was prepared to have the chemical composition shown in Table 1 using converter, ladle refining, and continuous casting methods, and the cast steel material (slab) was hot rolled into steel plates with thicknesses of 19 to 100 mm.

[0082] When manufacturing thick steel plates for welded structures, we optimized the manufacturing specifications for the plates, which relate to their mechanical properties. As a pre-training step, the predictive model was first trained using a regression model on the training data. This model then linked the actual manufacturing specifications for the steel plates, the actual toughness values ​​in the thickness direction, the actual welding specifications, and the actual occurrence of defects in the welds to create a weld defect occurrence prediction model. Using this predictive model, we searched for and obtained the manufacturing specifications for the thick steel plates using Bayesian optimization based on a Gaussian probability distribution, so that the estimated defect occurrence rate in the welds would be within the range of the target defect occurrence rate in the welds, and the toughness in the thickness direction would be improved. When searching for manufacturing conditions for steel plates, the target chemical composition of the steel and rolling conditions for the steel plates were set in advance so that the tensile properties, dimensions, and shape could be met, and the molten steel casting speed, cooling conditions during casting, soft reduction conditions, and electromagnetic stirring conditions were searched for. Then, steel plates were manufactured according to the manufacturing specifications obtained from the searched conditions.

[0083] JIS No. 4 tensile test pieces were taken from the 1 / 4 position of the plate thickness of each manufactured thick steel plate in the width direction perpendicular to the rolling direction, and tensile tests were conducted in accordance with the provisions of JIS Z 2241 to investigate the tensile properties.

[0084] Furthermore, V-notch Charpy impact test specimens were taken from the 1 / 4 position of the plate thickness of each manufactured thick steel plate in the rolling direction in accordance with the provisions of JIS Z 2242, and Charpy impact tests were carried out in accordance with the provisions of JIS Z 2242 to determine the absorbed energy at 0°C (vE0) and evaluate the toughness of the base material. The test values ​​are shown in the column for toughness in the rolling direction in Table 2.

[0085] Furthermore, V-notch Charpy impact test specimens were taken from the center of the thickness of each manufactured thick steel plate in the thickness direction, with the notch positioned at the center of the thickness. For plate thicknesses less than 55 mm, dummy steel was pressed onto both ends of the test specimen, and V-notch Charpy impact test specimens were fabricated in accordance with the provisions of JIS Z 2242. Charpy impact tests were conducted in accordance with the provisions of JIS Z 2242, and the absorbed energy at 0°C (vE0) was determined to evaluate the toughness of the base material. The test values ​​are shown in the column for toughness through the thickness in Table 2. Each vE0 in Table 2 represents the average value of three test pieces.

[0086] To evaluate the degree of element enrichment at the center of the plate thickness, test specimens were taken from the center of the plate width and thickness of the manufactured steel plate and processed so that the cross section was in the longitudinal and thickness directions of the steel plate. Evaluation was performed using an electron probe microanalyzer (EPMA). The field of view was centered at the center of the plate thickness, and the Mn concentration was measured within a 5 mm x 10 mm field of view in the plate thickness direction. After electrolytic polishing, the Mn concentration was measured at an acceleration voltage of 20 kV, with a beam shape of 20 μm length and 20 μm steps. Of the 250 x 500 points within the field of view, the locations where the Mn concentration was 1.2 times or more the Mn element analysis value of the molten steel (ladle) were defined as element-enriched areas (including inclusions), and the area ratio of the element-enriched areas was calculated for 10,250 points.

[0087] In addition, test plates for welded joints (500mm wide x 1000mm long) were taken from each manufactured thick steel plate, and welded joints were fabricated by submerged arc welding with the groove shape shown in Figure 1. The welding conditions, including the heat input and welding materials, were changed depending on the strength of the steel plate. To evaluate the cold cracking of the welded joints produced, the steel plate on the flange side was measured in accordance with JIS G 0901 (2010), and the maximum echo height was evaluated as follows: if it was less than the DL line, it was marked "absent," if it was between the DL line and the DM line, it was marked "○," if it was between the DM line and the DH line, it was marked "△," and if it was above the DH line, it was marked "×."

[0088] The results of the evaluation tests obtained above are shown in Table 2. Test Nos. 9, 10, and 11 (comparative examples) are test results for steel plates manufactured under casting conditions that predicted a toughness value in the thickness direction of less than 27 J. Test No. 12 (comparative example) was manufactured under casting conditions that predicted a toughness value in the thickness direction of 27 J or more with a reduction ratio (slab thickness / product thickness) of 2.5 or more, and then rolled to a thickness reduction ratio of 2.

[0089] In the invention examples of Test Nos. 1 to 8 and 13 to 27, all achieved a thickness direction toughness value of 27J or more. These steel plates were then submerged arc welded, and ultrasonic testing of the weld heat affected zone confirmed no weld defects. On the other hand, in the comparative examples of Test Nos. 9 to 12, thick steel plates with a thickness toughness value of less than 27J were submerged arc welded, and ultrasonic testing of the weld heat affected zone confirmed weld defects.

[0090] [Table 1]

[0091] [Table 2] [Explanation of symbols]

[0092] 1 Flange side steel plate 2. Web side steel plate 3 Welded parts 4. Backing 5 Center of plate thickness 6 Stress in the thickness direction of the flange side steel plate 7. Cold cracking

Claims

1. In mass%, C: 0.03-0.16%, Si: 0.50% or less, Mn: 0.8 to 3.0%, P: 0.015% or less, S: 0.0050% or less, Al: 0.005-0.100%, Ti: 0.004 to 0.030%, N: 0.0015 to 0.0065%; Optionally, Cu: 0.01 to 1.00%, Ni: 0.01 to 2.50%, Cr: 1.5% or less, Mo: 1.0% or less, Nb: 0.1% or less, V: 0.2% or less, REM: 0.02% or less, Mg: 0.005% or less, B: 0.005% or less, The method comprises a step of acquiring manufacturing specifications for a steel plate for a welded structure, the remaining component of which is Fe and unavoidable impurities, using teacher data including the results of manufacturing specifications for the steel plate, mechanical property values ​​of the steel plate, the results of welding specifications, and the results of defect occurrence in welds, based on a weld defect occurrence prediction model that has been trained in advance by machine learning, and acquiring manufacturing specifications for the steel plate so that the defect occurrence rate in welds falls within a desired value range; and a manufacturing step of manufacturing the steel plate using the manufacturing specifications for the steel plate acquired in the manufacturing specification acquisition step for the steel plate, The obtained steel plate for welded structures has an area ratio of Mn-enriched parts in the center of the plate thickness of 3% or less, a Charpy absorbed energy vE0 at the center of the plate thickness in the plate thickness direction at a test temperature of 0°C of 27 J or more, and a yield strength of 325 to 750 N / mm ２ , and a tensile strength of 490 to 930 N / mm ２ A method for manufacturing a thick steel plate for a welded structure, characterized in that the yield ratio is 90% or less. Here, the Mn-enriched portion is defined as a region in an analysis field including the central portion of the plate thickness where the Mn concentration of the steel plate, expressed in mass%, is 1.2 times or more the Mn concentration of the molten steel in the ladle, expressed in mass%, The desired value for the defect occurrence rate of a weld is determined by measuring the steel plate on the flange side of a welded joint in accordance with JIS G 0901 (2010) and setting the maximum echo height to be less than the DL line.

2. A method for generating a weld defect occurrence prediction model during the manufacture of a thick steel plate for a welded structure according to claim 1, comprising: A method for generating a weld defect occurrence prediction model, characterized by using training data including the results of manufacturing specifications for thick steel plates, mechanical property values ​​of thick steel plates, results of welding specifications, and results of defect occurrence in welds, to learn through machine learning and generate a prediction model for defect occurrence in welds.

Images