Assessment of the Hot-Cracking Susceptibility of Welded Joints of the 7CrMoVTiB10-10 Bainitic Steel Used in Heat Exchangers

Abstract

Bainitic steel containing approx. 2.25% Cr and 0.6 Mo with micro-additions of V, Ti, and B, designated as 7CrMoVTiB10-10 (T/P24), is one of the new construction materials used in new supercritical power units. The weldability of 7CrMoVTiB10-10 is defined as the hot-cracking susceptibility, it should be stated that the hot cracking in the welded joints of 7CrMoVTi10-10 is determined by phenomena occurring in the high-temperature brittleness range (HTBR). In this work, the HTBR is established for both the base material and welding conditions, taking into account the critical temperature-strain intensity (CST) and the critical strain speed (CSS). The HTBR for 7CrMoVTi10-10 is 122 °C wide and covers temperatures from 1394 °C to 1516 °C. Under imposed deformation conditions (typical during welding), the HTBR extends to a width of 293 °C towards lower temperatures, i.e., from 1516 °C to 1223 °C. The CSS = 0.83 1/s, the CST = 0.003 1/°C, and the R_f index = 0.12, which can be adopted as the criteria for the susceptibility of 7CrMoVTiB 10-10 to hot cracking under imposed deformation conditions. The hot cracking in 7CrMoVTiB10-10 occurs as a result of the loss of cohesion by the thin liquid layer crystallising between the growing weld crystals. Such cracks appear when the CSS or the CST is exceeded within the HTBR.

1. Introduction

The development and modernization of energy-generating systems, aimed at meeting climate-protection requirements and increasing the efficiency of electricity generation, depend on the availability of new construction materials and on the technology used for their processing [1].

During the construction of new supercritical power units (the term “supercritical” refers to the parameters of power units designed to operate at a steam temperature of 565–620 °C and at a pressure of 30 MPa), it was found that due to problems related to manufacturing, delivery times, and technological difficulties, e.g., problems involved in welding the P91 steel and in the heat treatment of large low-rigidity components (e.g., membrane walls), it was necessary to develop a new steel for the power industry that would not require such a complex welding and heat-treatment procedure. It was assumed that the recommended service temperature should be 650 °C, with a creep resistance R_z100000 of min. 100 MPa. The new steel was to be used mainly for the construction of membrane walls for power boilers operating within supercritical parameters. One such solution is a bainitic steel containing approx. 2.25% Cr and 0.6 Mo with micro-additions of V, Ti, and B, designated as 7CrMoVTiB10-10 (T/P24). The strengthening of this steel is related to the precipitation of chromium carbides and boron carbonitrides in the bainitic structure.

This steel has been used for the construction of membrane walls in a number of power boilers across Europe. Unfortunately, despite its good heat resistance and creep resistance, the steel was very problematic in terms of technological processing. The main difficulty involved in welding 7CrMoVTiB10-10 is its hot-cracking susceptibility during the crystallisation of the welded joint, both at the manufacturing stage and at the assembly stage. The hot cracking results from phenomena occurring in the high-temperature brittleness range (HTBR).

There is no precise definition of the high-temperature brittleness range, i.e., the range of temperatures during weld crystallisation within which the material has reduced ductility and is susceptible to hot cracking. According to Prokhorov [2], the upper limit of the HTBR should be considered to be the liquidus temperature, whereas its lower limit is near the solidus. Lancaster [3], in turn, assumes that there is a temperature at which crystals merge to create a cohesive, though not completely solidified, mass that has a certain strength. During further cooling, the material gains ductility. The HTBR is the range between the cohesion temperature and the ductility temperature [3]. A similar definition was adopted in [4], where the HTBR was understood as the difference between the nil strength temperature (NST) and the ductility recovery temperature (DRT) upon cooling [4].

In [5,6], the upper limit of the HTBR is defined as the temperature at which bridges form between crystals that are unable to bear plastic deformations, whereas the lower limit is understood as the temperature at which the metal is capable of deforming by transcrystalline slips (Figure 1).

Basic and industrial tests of weldability indicate that the final effect of the phenomena occurring in the HTBR is macro- and microcracking in the welded joints [7]. This cracking should be treated as irreversible damage occurring during crystallisation, i.e., the co-existence of the solid and liquid phase.

Alloys with a broad HTBR are more susceptible to hot cracking due to the long period of the existence of a thin liquid film in their interdendritic spaces. Irrespective of internal factors during crystallisation, i.e., crystallisation shrinkage and thermal contraction, a solid–liquid material is subject to stresses and strains. Stresses are not a critical factor that determines cracking, as the forces during crystallisation are comparable to the stresses that can be borne by the solid–liquid lattice [7,8].

There are also many theories concerning the influence of strain or the strain rate on the hot-cracking susceptibility of alloys [8,9,10]. These phenomena are not fully explained. Research aimed at determining phenomena occurring in the HTBR and cracking criteria in this temperature range was conducted in the second half of the 20th century. Studies in this area were undertaken, e.g., by Novikov [10], Siworth [11], and Zheng [12]. A broad analysis of theories concerning cracking in the liquid–solid state was performed by Suyitno [13], Eskin [14], and Katgerman [8].

In their works, Novikov and Novik claim that at low strain rates, the main deformation mechanism in a semi-solid body is grain boundary sliding. The load applied to a liquid–solid alloy will be accommodated by the displacement of grain boundaries, which are lubricated by a thin liquid layer [15].

Prokhorov also proposed a semi-solid deformation model [2]. Crystallisation cracking depends on three factors: the size of the HTBR, the plastic-strain capacity, and the strain growth rate (Figure 2). The hot-cracking susceptibility of a welded joint can be expressed as Δε_z, indicating the reserve of strain capacity:

Δε_z = pmin − (Δε_sk + Δε_k)

(1)

where p_min—minimum plastic strain capacity of the material, Δε_sk—strain caused by free shrinkage, and Δε_k—strain caused by joint shape change.

Thus, Prokhorov postulated that an increase in film thickness and a decrease in grain size increase resistance to crystallisation cracking, whereas any non-uniformity of grain size increases cracking susceptibility [2]. The main measure for hot-cracking susceptibility is alloy plasticity in the liquid–solid state. A crack will occur if the strain of a semi-solid alloy or a welded joint exceeds its plasticity [2,16,17].

In Lancaster’s work [3], it was claimed that a broad HTBR is characteristic of alloys with a small strength decrease in temperature, and, thus, such alloys are susceptible to hot cracking. Alloys with a narrow HTBR, in turn, are resistant to hot cracking because the plasticity and strength curves are close to each other.

Assuming that the highest strain accumulates in the superheat zone, Dodd [18] and Metz and Flemings [19] explain the increase in hot-cracking susceptibility by the segregation of low-melting components, which increases the time of the liquid film existence. The Pellini theory is the basis for the hot-tearing criterion proposed by Clyne and Davies [20]. The time spent by a metal in the mushy state is the basic criterion adopted by Fredriksson in [21]. They claim that the last stage of solidification is the most susceptible to hot cracking; however, a further decrease in the liquid fraction among dendrites can prevent cracking due to bridging between adjacent dendrites.

Other authors suggest that it is not the strain but the strain rate that is the critical parameter for crystallisation cracking. This is explained by the fact that the strain rate during solidification is limited by the minimum strain rate at which the material will fracture [2]. Another strain-rate-based hot-tearing criterion is proposed by Rappaz [22]. Additional theories explaining the hot-cracking phenomenon assume that failure takes place at a critical stress. The liquid surrounding the grain is considered as a factor increasing stress in the semi-solid body. In this theory, a liquid-filled crack is considered as a crack-initiation site [1,20,22].

Niyama [23] and Feurer [24] adopted a theory that assumes that the main cause of hot cracking is the hindered feeding of the solid phase by the liquid. Based on this theory, hot cracking will not occur as long as there is no lack of liquid feeding. The first two-phase model was developed by Rappaz–Drezet–Gremaud in 1999 [22]. Their criterion is formulated on the basis of after feeding, which is limited by the permeability of the mushy zone. At the solidification front, the permeability is high, but it drops deeper in the mushy zone. A pressure drop in the crystallisation zone is a function of this permeability and the strain rate. If the local pressure becomes lower than a critical pressure, a cavity appears that acts as a crack-initiation site [16,17]. A further development of the criterion, complemented by the plastic deformation of the solid phase and a cavity-growth criterion, was proposed by Braccini [25]. The model is based on two geometric models: one for columnar dendritic crystallisation and one for an equiaxed dendritic structure. It was found that the strain rate decreases with the increasing solid fraction. The basic mechanisms and conditions of hot cracking, according to Katgerman, are shown in Figure 3 [8].

A major factor affecting interdendritic (crystallisation) cracking is the characteristic of the structure forming during crystallisation. The influence of the characteristic of crystallisation on the alloy’s reserve of plasticity—and, hence, its susceptibility to interdendritic tearing and cracking—is shown in Figure 4 [6].

Welds that have a cellular structure during crystallisation are the most susceptible to cracking. Cracking in such welds is facilitated by the smooth surfaces of the grain boundaries, where a strong segregation of the low-melting components occurs. In the case of cellular–dendritic crystallisation, the specific surface area of the grain boundaries is larger, which results in a lower concentration of low-melting phases per unit of area and a lower susceptibility to tearing or cracking. When growing, the side branches of the dendrites become interlocked, which gives the alloy (or the weld) additional strength and reduces its cracking susceptibility [6].

Suyitno’s discussion of the theories describing the phenomena and mechanisms leading to hot cracking [13] and their comparison with casting practice also confirm that the theories developed to date do not take into account all the phenomena occurring during welded-joint crystallisation. The theories are mainly related to casting processes and do not take into account the heterogeneous conditions of crystallisation in the weld pool.

Suyitno confirmed that the theories of Feurer, Kategerman, Prokhorov, Rappaz–Drezent–Gremaud, and Braccini assume that hot cracking depends on the casting speed and indicate that the central part of the casting is susceptible to cracking. Experimental data confirm these correlations. However, all the theories, except for the Rappaz–Drezent–Gremaud theory, fail to indicate the significant influence of a reduced casting speed at the initial stage of the process, which is inconsistent with practical experience.

Despite extensive research that has led to the determination of hot-cracking criteria and the development of material-crystallisation models, this type of cracking still poses a major problem. There are no comprehensive studies concerning the hot-cracking phenomena and mechanisms in 7CrMoVTiB10-10, which has led to numerous failures and delays in the construction of supercritical power units.

Thus, assuming that the weldability of 7CrMoVTiB10-10 is defined as hot-cracking susceptibility, it should be stated that the hot cracking in the welded joints of 7CrMoVTi10-10 is determined by phenomena occurring in the HTBR. In this work, the HTBR is established both for the base material and for welding conditions, taking into account the critical temperature strain intensity (CST) and the critical strain speed (CSS).

2. Materials and Methods

The chemical composition and the mechanical properties of the material tested (steel 7CrMoVTiB10-10), according to the Vallourec & Mannesmann (V&M) certificate for heat 41037, are presented in

Table 1. Chemical composition and mechanical properties of 7CrMoVTiB10-10 (T/P 24).

Steel Grade	According to EN 10204:2004
T/P 24	%C	%Si	%Mn	%P	%S	%Al	%Cr	%Mo	%V
min	0.050	0.150	0.30	-	-	-	2.200	0.900	0.200
max	0.100	0.450	0.70	0.020	0.010	0.020	2.600	1.100	0.300
7CrMoVTiB10-10 bainitic steel T/P 24	Chemical composition according to the certificate—casting 41037
	0.074	0.283	0.55	0.014	0.001	0.014	2.422	0.986	0.249
	Mechanical properties according to the certificate—casting 41037
	Rp0,2 (MPa)		Rm (MPa)			A5 (%)		HV
	523		617			19.5		220.0

. For comparison, Table 1 also shows the required chemical composition according to the standard and selected mechanical properties required for this steel grade according to EN 10204:2004. The steel was delivered in heat-treated condition, i.e., following normalising, which ensured its homogeneous bainitic structure.

Based on an analysis of the chemical composition and the mechanical properties set out in the certificate for casting 41037, it was established that the test material met the requirements of EN 10204:2004 for 7CrMoVTiB10-10.

2.1. Methodology for Testing the Hot-Cracking Susceptibility of 7CrMoVTiB10-10

In order to determine the HTBR for the base material, i.e., 7CrMoVTiB10-10, it was necessary to determine the actual solidus and liquidus temperatures. The temperatures were measured by differential thermal analysis (DTA). The DTA was performed using a Setaram SETSYS thermal analyser. A TG–DTA head was used to measure the enthalpy of phase transitions upon heating 7CrMoVTiB10-10 up to 1700 °C and upon cooling down from that temperature. A cooling rate of 6 °C/min was applied. A type B (Pt-Rh 6%/Pt-Rh 30%) thermocouple was used for the DTA. The material was kept in an inert argon atmosphere (Ar 99.9999%), with a flow rate of 1.45 l/h. The temperatures of the beginning and the end of the transition were measured by the one set point method.

In order to assess the steel’s behaviour in the HTBR during welding, the heat cycle was determined, i.e., changes in the temperature of particular points upon cooling from a temperature close to the solidus. Knowledge of the heat cycle makes it possible to specify the area where structural changes occur in the material during welding.

The tests were conducted on cylindrical samples measuring Ø 6 × 120 mm, using a Gleeble 3800 thermal–mechanical simulator (Figure 5). Four type S thermocouples were applied to the samples: in the weld’s axis and at 2, 5, and 8 mm from the weld’s axis (Figure 6). After the samples were mounted in copper grips, maintaining a fixed grip distance of 33 mm, they were heated in an argon atmosphere at a rate of 200 °C/s, up to a temperature close to the solidus, and then freely cooled. During the experiment, changes in temperature were recorded at particular HAZ points. Based on the results, equations describing temperature change in time during sample cooling were determined. It was found that the curve describing the temperature change as a function of time during cooling was a polynomial of degree two. Based on an analysis of the regression and correlation of a function of one variable (non-linear), the determined relation was confirmed as true.

In order to determine the nil strength temperature (NST) upon heating for 7CrMoVTiB10-10, tests on cylindrical samples measuring Ø 6 × 90 mm were conducted on a Gleeble 3800 simulator. A type S thermocouple was applied to the samples, which were subsequently placed in copper grips in the chamber. A fixed grip distance of 52.4 mm was maintained. Following the removal of air from the chamber, it was filled with argon (up to 0.14 hPa). A minimum initial load of 0.6–0.7 kN was set and maintained until the end of the experiment. The samples were heated at a rate of 20 °C/s up to 400 °C and then at a rate of 1 °C/s. The NST was determined as the temperature at which a sample lost its cohesion.

The nil ductility temperature (NDT) and the ductility recovery temperature (DRT) were determined in order to assess the width of the HTBR and to determine the influence of the structure on the hot-cracking mechanism. The temperature at which a sample’s reduction in area is less than 5% was adopted as the NDT, and a reduction in area of more than 5% was adopted as the DRT. The value of reduction in area was adopted as the measure of plasticity, according to the following formula:

$$A=\frac{d_p-d_k}{d_k}\ast 100\%$$

(2)

where d_p—initial diameter, and d_k—final diameter.

The tests were conducted using a Gleeble 3800 simulator. Cylindrical samples measuring Ø 6 × 120 mm were mounted in copper grips in a protective argon atmosphere. In order to identify the NDT, the samples were heated to a pre-defined temperature within the HTBR, lower than the NST, annealed for 5 s, and then stretched at a defined fixed rate. The DRT was determined during the cooling of the samples from a temperature close to the NST down to a pre-defined test temperature and the subsequent stretching at a fixed rate. In each of the experiments—both NDT and DRT—two deformation rates were applied: 1 mm/s and 20 mm/s. Examples of changes in sample temperature and deformation as a function of time during the NDT test are shown in Figure 7.

The determined NDT and DRT and the mechanical properties of the steel were used to determine the HTBR, understood as the difference between the NST and the DRT. The results obtained made it possible to draw a ductility curve, understood as reduction in area, and a curve showing changes in strength as a function of temperature upon heating and cooling.

2.2. Determination of the Critical Strain Speed and Strain Intensity during Remelting under Imposed Deformation

The transvarestraint test was used to assess the HTBR under imposed deformation, characteristic of the welding process. The test consists in the rapid bending of flat samples over a cylindrical mandrel (Figure 8a). Samples are bent perpendicularly to the electric arc travel direction, in an argon atmosphere, using the TIG method (Figure 8b). The strain depends on the sample thickness and the radius of curvature of the mandrel. The test was conducted using mandrels with the following radius lengths, 50, 100, 150, 200, 224, and 324 mm, with samples having the following dimensions: 120 mm × 90 mm × 3.5 mm. The material was remelted with an alternating current of 80 A. The melting rate was 2 mm/s. The process parameters were selected to achieve a complete joint penetration.

The deformation value was determined according to the following formula:

$$\epsilon =\frac{g}{2R}\times 100\%$$

(3)

where ε—deformation value, (%); g—sample thickness, (mm); and R—mandrel radius of curvature, (mm).

Subsequently, the length of the longest crack in the weld axis (Lmax) and the total length of all cracks (Lmax) were determined. Based on length of the crack in the weld axis (Lmax) and the corresponding deformation, as well as the welding rate (v_s), the crack-growth time (tmax) was calculated using the following formula:

$$t_{\max }=\frac{L_{\max }}{v_s}$$

(4)

where t_max—crack-growth time, (s); L_max—longest crack, (mm); and v_s—welding rate, (mm/s).

The results of the calculations were used to draw a ductility curve ε = f(T) and to determine the HTBR under imposed deformation conditions. The high-temperature brittleness range was determined as the difference between the NST and the temperature of the end of the longest crack (T_k). The procedure and the methodology for determining the HTBR based on the results of the transvarestraint test are shown in Figure 9.

2.3. Methodology of the Metallographic Examinations

In order to explore the hot-cracking mechanism, the HTBR tests were complemented with metallographic examinations. The structural examinations were conducted on fracture surfaces after the Gleeble 3800 simulator tests and on samples cut out perpendicularly to the remelting direction applied in the transvarestraint test. The metallographic sections were etched in a 5% solution of nitric acid (nital). Observations at magnifications of up to 50x were performed in the dark field on an OLYMPUS SZX 9 stereoscopic microscope (SM), whereas structural examinations at larger magnifications were conducted in the bright field on an Olympus GX71 light microscope (LM), as well as using the secondary electron (SE) and the back-scattered electron (BSE) techniques, on a Hitachi S 3400N scanning electron microscope (SEM). SE images provide a good representation of the surface topography, whereas BSE images indicate differences in the chemical composition. The structural evaluation was complemented by an EDS microanalysis of the chemical composition, performed on a SEM microscope equipped with a Thermo Noran microanalyser.

3. Results

3.1. DTA Results

An analysis of the DTA curves both during sample cooling (Figure 10a) and during sample heating (Figure 10b) made it possible to determine the liquidus upon heating and the solidus upon cooling. These temperatures are particularly important in the welding process, where the material melts in front of the heat source and crystallises after the passing of the electric arc. The results are presented in

Table 2. Solidus and liquidus temperatures for the 7CrMoVTiB10-10 steel grade.

Alloy Name	Solidus on Cooling, °C, Tpeak	Liquidus on Heating, °C, Tpeak
7CrMoVTiB10-10 bainitic steel	1515	1574

.

3.2. Assessment of the Heat Cycle

Heat cycle simulations conducted on a Gleeble 3800 simulator (Figure 11a) enabled the determination of the change in temperature in time (Figure 11b). It was assumed that the determined heat cycle corresponded to the welding heat cycle and could be used to assess the HTBR under remelting conditions.

It was found that the curve describing the temperature change as a function of time during cooling was a polynomial of degree two. Based on an analysis of the regression and correlation of a function of one variable (non-linear), it was found that the determined relation was significant.

3.3. Determination of the Nil Strength Temperature (NST)

Table 3. Results of the NST measurement, °C.

NST, °C
Measurement	1	2	3	4	5	6
7CrMoVTiB10-10 bainitic steel	1506	1537	1515	1508	1532	1518

shows the results of the measurements of the nil strength temperature (NST) upon heating. Figure 12 shows the fracture surface and examples of the crack area microstructure revealed on a section perpendicular to the fracture surface. The average temperature at which cracks appeared in the six samples was adopted as the NST. The temperature was 1516 °C.

Based on the results of the metallographic examinations, it was found that the loss of strength by 7CrMoVTiB10-10 at high temperatures occurred as a result of the partial melting of primary structure grains and the splitting of liquid along primary austenite boundaries (Figure 12).

3.4. Determination of the Nil Ductility Temperature (NDT), the Ductility Recovery Temperature (DRT), the High-Temperature Brittleness Range (HTBR), and the Cracking-Susceptibility Index Rf

The results of the Gleeble 3800 simulations, aimed at determining the nil ductility temperature (NDT) (Figure 13), the ductility recovery temperature (DRT) (Figure 14), the high-temperature brittleness range (HTBR), and the cracking-susceptibility index R_f, are shown in

Table 4. Results of the measurement of the temperatures characteristic of the crystallisation process of 7CrMoVTi 10-10.

Vo (mm/s)	NDT (°C)	NST (°C)	DRT (°C)	HTBR (°C)	ΔHTBR (°C)	Rf
1	1400	1516	1390	1390–1413	126	0.11
20	1410	1516	1394	1394–1413	122	0.12

and Figure 15, whereas the results of the metallographic and fractographic examinations are shown in Figure 16 (for the NDT test) and in Figure 17. An analysis of the fracture surfaces after the NDT test at 1360 °C indicates that grains undergo considerable deformation and lose cohesion as a result of the decrease in the surface area (Figure 16c). The partially melted grains on the surface of the sample subjected to deformation at 1360 °C result from the overheating of the material in the final phase of decohesion, due to the reduced surface area and the resistance heating (Figure 16d). Deformation at 1380 °C leads to a much smaller reduction in area (approx. 17%) (Figure 15a and Figure 16a).

Significantly deformed grains and numerous side cracks (cracks propagating from the main crack) along the grain boundaries were observed on the fracture surface. This indicates a change in the mechanism of cracking, which occurs as a result of the melting of the primary structure grains. This is confirmed by the intercrystalline fracture in the sample subjected to deformation at 1400 °C (Figure 16a). Moreover, thin layers of solidified liquid at grain boundary interfaces can be seen in Figure 16b. Thus, at high temperatures, 7CrMoVTiB10-10 fails as a result of the partial melting of primary structure grains and the subsequent rupture of the liquid film along the grain boundaries.

The results of the DRT test indicated that during cooling from a temperature close to the NST, irrespective of the strain speed, the DRT was similar: 1390 °C for a strain speed of 1 mm/s and 1394 °C for a strain speed of 20 mm/s (Figure 15).

Metallographic examinations of the fracture surfaces after the DRT test indicate that the loss of cohesion by the alloy occurs, just as in the case of sample deformation during heating, due to the rupture of the liquid at the boundaries of partially melted grains (Figure 17a). As the temperature drops, the grains in the material’s decohesion area are subjected to plastic deformation according to the direction of the tensile force (Figure 17c). At 1380 °C, side cracks along the boundaries of grains with strongly deformed surfaces were revealed (Figure 17d). The reduction in area observed at 1360 °C results in the partial melting of grains due to the resistance heating, and the resultant intercrystalline liquid fills the side cracks (Figure 17d).

Based on the results of the Gleeble 3800 simulations of the welding process, the HTBR was determined for the 7CrMoVTiB 10-10 steel. At a strain speed of 1 mm/s, the HTBR is 126 °C wide and covers temperatures from 1390 °C to 1516 °C (Table 4). The HTBR for the strain speed of 20 mm/s, typical of welding processes, is similar, covering a width of 122 °C and temperatures from 1394 °C to 1516 °C. Thus, no significant influence of the strain speed on the HTBR under uniaxial deformation at high temperatures was identified (for the DRT test).

The NDT tests upon heating enabled the determination of the NDT for various strain speeds. For 7CrMoVTiB10-10, the NDT for a strain speed of 1 mm/s is 1400 °C and for a strain speed of 20 mm/s is 1410 °C (Table 4). The hot-cracking-susceptibility index (R_f) is similar for each strain speed and ranges from 0.11 to 0.12.

3.5. Determination of the Critical Strain Speed (CSS), the Critical Temperature Strain Intensity (CST), and the HTBR during Remelting under Imposed Deformation by Means of the Transvarestraint Test

Figure 18 shows examples of samples after bending during remelting in the transvarestraint test, whereas

Table 5. Results of the transvarestraint test for 7CrMoVTiB10-10.

Parameter	Deformation ε = g/2R∗100 (%)
	3.5	1.75	1.16	0.87	0.78
Sum of crack lengths ΣLi (mm)	7.90	4.90	5.52	6.02	3.47
Max crack length Lmax (mm)	1.80	1.50	1.30	1.28	1.20
Sample thickness (mm)	3.5	3.5	3.5	3.5	3.5
ΣN—number of cracks	7	7	6	7	3
Welding rate (mm/s)	2	2	2	2	2
tmax (s)	0.9	0.75	0.65	0.64	0.6

presents the results of the calculations of the lengths of identified cracks and the times of their growth. The results of these calculations enabled the determination of the crack length and crack-growth time as a function of deformation (Figure 19a). The results obtained made it possible to determine a logarithmic curve of the crack-growth time as a function of deformation t_max = f(ε) (Figure 19b). The results of the calculations were used to draw a ductility curve ε = f(T) (Figure 20) and to determine the HTBR under imposed deformation conditions.

The HTBR was determined as the difference between the NST and the temperature of the end of the longest crack (T_k). The results of the metallographic examinations on a flat cross-section transverse to the remelting direction and the results of the fractographic examinations are shown in Figure 21. Figure 21 presents the results of the microanalysis of the particles revealed on the fracture surface.

Based on an analysis of the regression and correlation of a function of one variable (non-linear), the determined relations were confirmed as true. Moreover, the critical strain speed (CSS), understood as the tangent of the inclination angle between the tangent to the crack-growth curve and the crack-growth-time axis (Figure 19), and the critical temperature strain intensity (CST), being the tangent of the angle between the tangent to the ductility curve ε = f(T) and the temperature axis (Figure 20), were determined. The results of the calculations are presented in

Table 6. Results of the assessment of the HTBR for 7CrMoVTiB 10-10 under imposed deformation.

Alloy	CSS, 1/s	CST, 1/°C	ΔHTBR °C	HTBR, °C
7CrMoVTiB 10-10	0.83	0.003	293	1223 ÷ 1516

.

An assessment of the hot-cracking susceptibility of 7CrMoVTiB10-10 under imposed deformation conditions was performed. Such conditions are common during the manufacture of large-sized boiler structures, for which the steel examined is intended. The critical temperature strain intensity (CST), which is is 0.003 1/°C for the steel in question (Figure 20), the critical strain speed (CSS)—0.83 1/s (Figure 19b), and the brittleness threshold ε_p, i.e., the deformation at which no cracking occurs (<0.55%), were adopted as the cracking criteria (Table 6).

The results obtained in the transvarestraint test made it possible to determine the crack-growth time as a function of deformation (Figure 19b) and the ductility curve ε = f(T) (Figure 20). An analysis of these curves enables the determination of the HTBR. The width of this range determines the possibility of hot cracking occurring during welding. For the steel in question, the HTBR under imposed deformation conditions is 293 °C wide and covers temperatures from 1223 °C to 1516 °C (Figure 20, Table 6).

Analysis of the microstructure on the surface perpendicular to the remelting direction revealed a structure typical of a welded joint, i.e., an area of the base material having a martensitic–bainitic structure (Figure 21a), the heat-affected zone, and the fusion zone, characterised by an orthogonal arrangement of the primary austenite crystals (Figure 21b). Numerous cracks, propagating both from the weld face surface (Figure 18) and in the root (Figure 21a), were identified in the fusion zone. The cracks develop along the boundaries of primary austenite crystals, which indicates that they occur during the crystallisation of the fusion zone.

The examinations of the fusion zone’s fracture surface confirm that the cracks grow along the boundaries of primary austenite grains (Figure 21c,d). The cracks identified are hot in nature, as indicated by the side cracks along the grain boundaries and the non-fitting of particular grains (Figure 21d). The cracks occur as a result of the partial melting of crystals and the rupture of the thin liquid film in interdendritic spaces. The areas with the characteristic ductile–brittle fracture formed as a result of the breaking of the fusion zone during the preparation of the sample for the tests (Figure 21c). In these areas, inclusions of non-metallic phases having a characteristic cuboid shape were identified (Figure 22).

The microanalysis of the chemical composition, performed by the EDS method, indicated that the precipitates were rich in nitrogen and titanium—thus, they were titanium nitrides (Figure 22). Microanalysis of the chemical composition in the area of the “bridge” that spans the two crack edges revealed elements typical of 7CrMoVTiB10-10, which indicates that the bridges form as a result of the solidification of the liquid coming from the partially melted crystals. This confirms that the main cracking mechanism in 7CrMoVTiB10-10 under imposed deformation, typical of welding processes, is hot cracking.

4. Discussion of the Results

An analysis of the literature data indicates unambiguously that a steel suitable for application in the membrane walls of supercritical boilers, which at the same time does not require heat treatment after welding, needs to be developed and applied in industrial practice. One such solution was the use of the bainitic steel designated as 7CrMoVTiB 10-10 (TP24). However, despite the fact that the steel has very good mechanical properties, including creep resistance up to 650 °C, it was found that the possibility of its application was determined by its technological properties. The main technological problems related to the production of membrane walls of 7CrMoVTiB 10-10 include the susceptibility of welds to cracking during welding. Despite extensive research into this phenomenon, the causes of welded-joint cracking have not been unambiguously explained. These technological difficulties have led to the limited use of 7CrMoVTiB 10-10 in supercritical boilers.

The first stage of the research described in this paper, aimed at exploring the phenomena determining the cracking of welded joints of 7CrMoVTiB 10-10, was the determination of the hot-cracking susceptibility of this steel grade. The solidus and liquidus temperatures were identified by DTA (Table 2), and the welding heat cycle was determined (Figure 11). This information corresponded to the manufacturer’s data provided in [1]. On this basis, a thermal–mechanical simulator was used to determine the temperature upon heating, at which the material’s strength became nil (NST) (Table 4), and the temperature upon cooling, at which the material recovered its ductility (DRT) (Table 4). These data made it possible to identify the HTBR for the base material, i.e., 7CrMoVTiB 10-10 (Table 4). The HTBR for this steel is 126 °C wide for a strain speed of 1 mm/s and 122 °C wide for a strain speed of 20 mm/s. Within this temperature range, the steel is characterised by a considerably lower resistance to hot cracking. The assessment of the fracture surfaces for failures that occurred within this temperature range unambiguously indicated the hot-cracking mechanism (Figure 12c). The Rf index, which is 0.11–0.12 (Table 6) for 7CrMoVTiB 10-10, can be adopted as the hot-cracking-susceptibility criterion. When the material is heated, the edges of austenite crystals partially melt, and, as the liquid loses cohesion along the grain boundaries, local microcracks form and coalesce, leading to sample failure.

Welding is a dynamical process in which the electric arc (heat stream) moves with time, leading to the rapid melting of the material, and subsequently the weld pool crystallises. Changes in the position of the electric arc lead to a temperature gradient and, thus, variable stresses and strains. These factors result in the increased hot-cracking susceptibility of a welded joint. It is very important, especially for a joining process, to identify the HTBR width under imposed deformation during welding and the cracking criteria. The test consisting of remelting the sample under deformation (transvarestraint test) enabled the determination of the HTBR for the welding process—293 °C (Figure 20). The critical strain speed, which is 0.83 1/s, and the critical temperature strain intensity, which is 0.003 1/°C (Table 6), can be adopted as the cracking criteria. The data obtained make it possible to describe the hot-cracking mechanism during the welding of 7CrMoVTiB 10-10, as presented in Figure 23.

Analysis of the results of the metallographic and fractographic examinations indicates that during welding, hot cracks form in the weld as a network of fine interdendritic cracks. Those microcracks occur within the HTBR as a result of nil ductility, as described in Prokhorov [2] and Nivkov’s [10] works. The cracking mechanism is related to the rupture of the bridges between adjoining dendrites of the crystallising weld (Figure 21c) and the subsequent loss of cohesion by the intergranular liquid. At the next stage, the microcracks coalesce into a main crack as a result of the strains involved in weld crystallisation and the influence of the welding heat cycle (Figure 21).

The hot-cracking phenomenon and the mechanism in the welded joints of 7CrMoVTiB10-10, described in this work, significantly complement the knowledge in the area of materials engineering but are also significant due to the described phenomenon of hot cracking in the welded joints of 7CrMoVTiB10-10 during submerged arc welding. Exploring these mechanisms and determining cracking criteria make it possible to develop a welding technology that ensures that the membrane walls of supercritical boilers are free of welding defects, including the particularly dangerous hot microcracks. Such cracks can lead to a failure of a power system, both during its assembly and during its operation.

5. Conclusions

The following conclusions were formulated based on the tests and examinations conducted and the analysis of their results.

• The determined temperatures characteristic of the crystallisation area, i.e., the liquidus—1574 °C, the solidus—1515 °C, the NST—1516 °C, the NDT upon heating—1400 °C, and the DRT upon cooling—1390 °C, as well as the determined heat cycle, make it possible to describe the phenomena occurring at high temperatures during weld crystallisation and to determine the high-temperature brittleness range. The HTBR for 7CrMoVTi10-10 is 122 °C wide and covers temperatures from 1394 °C to 1516 °C.
• Under imposed deformation conditions (typical during welding), the HTBR extends to a width of 293 °C towards lower temperatures, i.e., from 1516 °C to 1223 °C.
• The critical strain speed (CSS) = 0.83 1/s, the critical temperature strain intensity CST = 0.003 1/°C, and the Rf index = 0.12 can be adopted as the criteria for the susceptibility of 7CrMoVTiB 10-10 to hot cracking under imposed deformation conditions.
• Hot cracking in 7CrMoVTiB10-10 occurs as a result of the loss of cohesion by the thin liquid layer crystallising between the growing weld crystals. Such cracks appear when the CSS or the CST is exceeded within the HTBR.