Failure Analysis on a Collapsed Flat Cover of an Adjustable Ballast Tank Used in Deep-Sea Submersibles

applied sciences

Article Failure Analysis on a Collapsed Flat Cover of an Adjustable Ballast Tank Used in Deep-Sea Submersibles

Fang Wang 1, Mian Wu 2, Genqi Tian 3, Zhe Jiang 1, Shun Zhang 1, Jian Zhang 2 and Weicheng Cui 1,4,*

1 Shanghai Engineering Research Center of Hadal Science and Technology, Shanghai Ocean University, Shanghai 201306, China; [email protected] (F.W.); [email protected] (Z.J.); Shun_zhang@rainbowﬁsh11000.com.cn (S.Z.) 2 Jiangsu Provincial Key Laboratory of Advanced Manufacture and Process for Marine Mechanical Equipment, Jiangsu University of Science and Technology, Zhenjiang 212003, China; [email protected] (M.W.); [email protected] (J.Z.) 3 School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China; [email protected] 4 School of Engineering, Westlake University, Hangzhou 310024, China * Correspondence: [email protected]

Received: 14 October 2019; Accepted: 28 November 2019; Published: 3 December 2019

Abstract: A flat cover of an adjustable ballast tank made of high-strength maraging steel used in deep-sea submersibles collapsed during the loading process of external pressure in the high-pressure chamber. The pressure was high, which was the trigger of the collapse, but still considerably below the design limit of the adjustable ballast tank. The failure may have been caused by material properties that may be defective, the possible stress concentration resulting from design/processing, or inappropriate installation method. The present paper focuses on the visual inspections of the material inhomogeneity, ultimate cause of the collapse of the flat cover in pressure testing, and finite element analysis. Special attention is paid to the toughness characteristics of the material. The present study demonstrates the importance of material selection for engineering components based on the comprehensive properties of the materials.

Keywords: ﬂat cover; adjustable ballast tank; fracture; high-strength maraging steel; toughness

1. Introduction Deep-sea manned/unmanned submersibles are the necessary high-tech equipment for ocean exploration. It is used to carry crews or equipment to various deep-sea complex environments for efficient exploration, scientific investigation and resource exploitation [1,2]. Deep-sea submersibles contain multiple complex systems. Buoyancy regulation system is one of the important subsystems of the submersible. The adjustable ballast tank is one of the key components, which can ensure that the submersible has a good ability of depth setting and weight fine-tuning [3–5]. Through the high-pressure seawater pump, seawater is pumped in and out of the ballast tank, and the buoyancy balance of the submersible in seawater is adjusted within a certain range. Reducing the weight for better performances such as range, speed, and payload is a significant consideration for the designers of the submersibles, as the weight-to-displacement ratio is used to evaluate the structural efficiency of underwater vehicles [1,2]. The weight of the submersible is distributed among the main components. At present, a spherical adjustable ballast tank with its promising application to a sea depth of 11,000 m and a volume of 300 L is designed, where domestic

Appl. Sci. 2019, 9, 5258; doi:10.3390/app9235258 www.mdpi.com/journal/applsci Appl. Sci. 2019, 9, x FOR PEER REVIEW 2 of 17

Reducing the weight for better performances such as range, speed, and payload is a significant consideration for the designers of the submersibles, as the weight-to-displacement ratio is used to

Appl.evaluate Sci. 2019, 9 the, 5258 structural efficiency of underwater vehicles [1,2]. The weight of the submersible2 of 15 is distributed among the main components. At present, a spherical adjustable ballast tank with its promising application to a sea depth of 11,000 m and a volume of 300 L is designed, where domestic ultra-highultra-high strength strength maraging maraging steel steel 18Ni(350) 18Ni(350) was was used used for for first first time time for for this this purpose, purpose as, shownas shown in in FigureFigure1. The 1. adjustable The adjustable ballast ballasttank is composed tank is composed of upper and of upper lower hemispheres, and lower hemispheres, which are connected which are by aconnected Huff clamp. by a An Huff O-ring clamp. seals An theO-ring hemispheres seals the hemispheres at the end. Fine-tuningat the end. Fine makes-tuning the makes weight the of weight the submersibleof the submersible maintain positive maintain buoyancy positive or buoyancy produce or enough produce negative enough buoyancy negative for buoyancy safely sitting for safely at bottom.sitting A at flat bottom. cover is A beneficial flat cover for is installationbeneficial for of installation different cabin of different piercing cabin parts. piercing Sealing isparts. carried Sealing out is throughcarried a radialout through O-ring a and radi isal evenly O-ring locked and is withevenly the locked sphere with by screws.the sphere The by selection screws. ofThe maraging selection of steelmaraging for the current steel for ballast the current tank is ballast exactly tank based is exactly on the principlebased on ofthe strength principle enhancement of strength enhancement for lower weight.for lower This is weight. also the This first timeis also that the this first material time that has beenthis material used in deep-sea has been pressure used in vessels. deep-sea Some pressure of thevessels. candidate Some materials of the candidate for underwater materials pressure for underwater hulls, such pressure as titanium hulls, and such high as titanium strength and steel, high andstrength their main steel, properties and their can main be foundproperties in literature can be found [6–9]. in The literature candidate [6– material9]. The candidate for pressure material hulls for usingpressure 18Ni grade hulls using maraging 18Ni steels grade has ma beenraging preliminarily steels has been investigated preliminarily in terms investigated of their in application terms of their history,application performance, history, and performance, manufacturing and capability manufacturing by the authors capability [9]. by The the damage authors tolerance [9]. The related damage performancestolerance related of 18Ni performances grade maraging of 18Ni steels, grade including maraging yield steels, ratio ( σ includingy/σb) and yield fracture ratio toughness, (σy/σb) and havefracture been evaluated. toughness, There have are been basically evaluated. four wrought There are commercial basically four maraging wrought steels commercial of the 18 percent maraging nickelsteels family of the i.e., 18 18Ni percent (200), nickel 18Ni family (250), i.e., 18Ni 18Ni (300), (200) and, 18Ni 18Ni (250), (350) 18Ni with (300), yield strengthand 18Ni ranging (350) with from yield 1400strength MPa to 2400ranging MPa from [10]. 1400 In the MPa development to 2400 MPa of ‘MIR’ [10]. In submersibles the development in the 1980s of ‘MIR [11’], submersibles new techniques in the to produce1980s [11], high new strength, techniques high to nickel-content produce high strength, steel 18Ni(250) high nickel for pressure-content hullssteel 18Ni(250) are applied for for pressure its twohulls pressure are applied spheres. for Up its to two now, pressure other grades spheres. of 18Ni Up to series now, of other maraging grades steels of 18Ni have series no application of maraging experiencesteels have in deep-sea no application pressure experience hulls. in deep-sea pressure hulls.

(b) (a)

Figure 1. Sketch of the adjustable ballast tank and its sealing cover: (a) front view of the adjustable ballastFigure tank; 1. ( bSketch) description of the adjustable of the sealing ballast cover. tank and its sealing cover: (a) front view of the adjustable ballast tank; (b) description of the sealing cover. The increase in strength is very powerful in reducing the weight of pressure hulls, but the only concern isThe that increase increased in strength strength generallyis very powerful leads to in a decreasereducing in the toughness weight of [12 pressure,13] while hulls, it is generallybut the only notconcern considered is that in the increased design of strength deep-sea generally compressive leads components. to a decrease Plane in toughness strain fracture [12,13] toughness while it is is angenerally important not factor considered to represent in the the design crack of resistance deep-sea property compressive of the components. material. Wherein, Plane strain 18Ni(250) fracture withtoughness yield strength is an important of 1700 MPa factor has to the represent plane-strain the crack fracture resistance toughness property level of of the 85–110 material. MPa Wherein,m1/2, · but18Ni(250) the value ofwith 18Ni(350) yield strength with yield of strength1700 MPa of 2400has the MPa plane rapidly-strain reduces fracture to 30–50 toughness MPa m le1/2vel[9 ,12of ,8513].–110 · ImprovingMPa·m the1/2, buttoughness the value and plasticity of 18Ni(350) of maraging with yield steel strength can be done of 2400 in various MPa ways, rapidly such reduces as reducing to 30–50 harmfulMPa· elementsm1/2 [9,12 or,13]. gas Improving content in the steel toughness by a double and vacuum plasticity smelting of maraging process, steel controlling can be done inclusions in various morphology,ways, such adjusting as reducing microstructure harmful by elements special processing, or gas content and heat in steel treatment by a double processes vacuum [14]. Results smelting 5 haveprocess, shown thatcontrolling when the inclusion harmfuls elementsmorphology, in 18Ni(350) adjusting maraging microstructure steel are by reduced special to processing, 10− magnitude and heat at thetreatment same time, processes the number [14]. Results and volume have shown of inclusions that when are the greatly harmful reduced, elements which in 18Ni is an(350 important) maraging reasonsteel for are the red remarkableuced to 10−5 magnitude improvement at the of same fracture time, toughness the number of and ultra-pure volume 18Ni(350)of inclusions maraging are greatly steel. The influence of the toughness on the resistance to cracking for deep-sea components needs to be examined. The hydraulic pressure test on the adjustable ballast tank with its sealing flat cover was carried out to examine its performance. The ultimate goal of pressure test was to check whether it can Appl. Sci. 2019, 9, 5258 3 of 15 endure the external pressure of 126.5 MPa (i.e., 115 MPa for 11,000 m deep-sea environments times a safety factor of 1.1 for pressure testing. Note, the actual pressure value for the 11,000 m deep-sea environment is 113.8 MPa and this value has been used in the design calculation, but here for the test, a more conservative value of 115 MPa is used). There are sensor mounting holes and high pressure pipe mounting holes on the sealing cover. However, the flat cover collapsed during the loading process of external pressure in the high-pressure chamber. The pressure was high, which was the trigger of the collapse, but still considerably below the design limit of the hull. The failure can be caused by unexpected defective material properties, the possible stress concentration resulting from design/processing, or inappropriate installation method. In this paper, the mechanical properties of the actual materials used in the collapsed flat cover are re-examined by sampling and testing of the broken parts. Non-metallic inclusions analysis, micro-structure analysis, and fracture surface analysis are conducted to acquire the possible fracture cause. Furthermore, finite element analysis based on fracture mechanics is conducted to understand the ultimate cause of destruction. The analysis results demonstrate the importance of material selection for engineering components based on the comprehensive properties of the materials.

2. Description of the Material and Structure

2.1. Chemical Composition and Material Properties The preparation process of the material consists of vacuum induction furnace smelting and vacuum consumable furnace remelting. Solution heat treatment was conducted at 820 ◦C and kept for 1 h; aging treatment was conducted at 510 ◦C and kept for 4 h. Chemical composition and mechanical properties of 18Ni (350) material are tested after sequential steps of melting, melt treatment practice, forging process, heat treatments, machining operations, and inspection, which is the same way as the cabin shell material was treated. Their chemical compositions and mechanical properties are given in Tables1 and2 and the stress–strain curve is depicted in Figure2. This material has high tensile strength, the yield strength of which can reach about 2258 MPa and there is no obvious yield platform. However, its elongation is only 5.4%. The current design code for the spherical hull used in deep-sea submersibles and the design guideline for its flat cover have not specified the requirements for elongation of the material. The designers of the current spherical adjustable tank select 18Ni (350) instead of traditional material considering its high strength to reduce the weight of the tank, which will benefit the capability of the whole submersible.

Table 1. Chemical composition of 18Ni(350), (wt.%).

Ni Co Mo Ti Al C Si Mn P S 18.80 11.72 4.41 1.32 0.125 0.006 0.02 0.01 0.004 0.002

Table 2. Mechanical properties of 18Ni(350).

σy, (MPa) σb, (MPa) A, (%) Z, (%) E, (MPa) υ KV at 20 ◦C, (J) 2258 2324 5.4 31.0 188 0.3 13.5

where σy: Yield strength; σb: Ultimate tensile strength; E: Young’s modulus; υ: Poisson’s ratio; A (%): Total elongation; Z (%): Area reduction. Appl. Sci. 2019, 9, x FOR PEER REVIEW 4 of 17 Appl. Sci. 2019, 9, x FOR PEER REVIEW 4 of 17

where σy: Yield strength; σb: Ultimate tensile strength; E: Young’s modulus; υ: Poisson’s ratio; A (%): whereAppl.Total Sci. eσlongation;2019y: Yield, 9, 5258 strength; Z (%): Areaσb: Ultimate reduction. tensile strength; E: Young’s modulus; υ: Poisson’s ratio; A4 (%): of 15 Total elongation; Z (%): Area reduction.

FigureFigure 2.2. Stress–strainStress–strain curve of 18Ni(350)18Ni(350) (solution(solution heat treatment:treatment: 820 °CC for 1 h; aging treatment: Figure 2. Stress–strain curve of 18Ni(350) (solution heat treatment: 820 °C◦ for 1 h; aging treatment: 510510 °CC for 4 h).h). 510 °C◦ for 4 h). 2.2.2.2. Design and GeometryGeometry of the Flat Cover 2.2. Design and Geometry of the Flat Cover TheThe opening opening of of the the ballast ballast tank tank is sealed is sealed by flange by connection flange connection through thethrough flat cover. the flat The calculationcover. The ofcalculation theThe thickness opening of the of thickness ofthe the flat ballast cover of the is tankflat carried cover is sealed out is carried according by flange out toaccording the connection requirements to the through requirement of the the circular flats of cover.the flat circular cover The calculationspecifiedflat cover in specified of the the guidelines thickness in the guid ofof steeltheelines flat pressure coverof steel is vessels carriedpressure [15 out v].essels according [15]. to the requirements of the circular flat cover specified in the guidelines of steel pressure vessels [15]. s k Pkc P = D ,c (1) δp = cD kt P , p c [σ] φ ct (1)  p = Dc    , (1)   t wherewhere DDcc= = 176176 mmmm isis thethe calculatedcalculated diameterdiameter ofof thethe flatflat cover;cover; kk == 0.25 is the structural characteristic where Dc = 176 mm is the calculated diameter of the flat cover; k = 0.25 is the structural characteristic coecoefficient;fficient; PPcc isis thethe calculatedcalculated pressure,pressure, whichwhich can be 1.4 timestimes thethe designeddesigned working pressure, i.e., coefficient; Pc is the calculatedt pressure, which can be 1.4 times the designed working pressure, i.e., 1.41.4 × 113 == 158.2 MPa; [σ]t = σσyy/1.5./1.5. In In order order to to make make the the strength strength safety safety factor factor larger, the yield stress × t 1.4appliedapplied × 113 by by= 158.2 thethe designer designer MPa; [σ presently ]presently = σy/1.5. is Inis reduced reducedorder to to tomake a lowera lowe the level, rstrength level, 1900 1900 MPa,safety MPa, then factor then [σ ]larger,t [=σ]1266.7t = 1266.7theMPa; yield MPa;φ stress= ϕ1.0 = applied by the designer presently is reduced to a lower level, 1900 MPa, then [σ]t = 1266.7 MPa; ϕ = is1.0 the is the welding welding coe fficoefficient.cient. This This value value of ofδp δ=p =31 31 mm mm is is obtained obtained in in calculation calculation butbut waswas 3232 mmmm inin 1.0practice,practice, is the includingweldingincluding coefficient. thethe thicknessthickness This ofof value 1212 mmmm of in inδp the the= 31 partpart mm ofof is ΦΦ190 obtained190 mmmm andandin calculation thethe thicknessthickness but ofof was 2020 mm32mm mm inin thethein practice,partpart ofof ΦΦ1 including15050 mm,mm, asasthe shownshown thickness inin FigureFigure of 12 3mm 3.. Necessary Necessary in the part verification verification of Φ190 mm has has and been been the conducted conductedthickness byof by 20 designers designers mm in the to to partcheckcheck of the theΦ1 strength50strength mm, as andand shown stistiffnessffness in Figure ofof thethe structure3.structure Necessary andand verification itit isis consideredconsidered has tobeen to meet meet conducted the the requirements. requirements. by designers to check the strength and stiffness of the structure and it is considered to meet the requirements.

Figure 3. Dimension of the ﬂat cover (unit: mm). Figure 3. Dimension of the flat cover (unit: mm). Figure 3. Dimension of the flat cover (unit: mm). 2.3. Failure of the Flat Cover During Testing 2.3. Failure of the Flat Cover During Testing 2.3. FailureThe spherical of the Flat adjustable Cover During tank Testing with its sealing cover was put into a high pressure chamber to examine its strength by pressure test before application. The pressure test was conducted in the high pressure chamber of the HAST laboratory at Shanghai Ocean University. There was no real-time video monitoring equipment in the pressure chamber and the test was carried out in several steps of cyclic Appl. Sci. 2019, 9, x FOR PEER REVIEW 5 of 17

The spherical adjustable tank with its sealing cover was put into a high pressure chamber to examine its strength by pressure test before application. The pressure test was conducted in the high pressureAppl. Sci. chamber2019, 9, 5258 of the HAST laboratory at Shanghai Ocean University. There was no real-time5 of 15 video monitoring equipment in the pressure chamber and the test was carried out in several steps of cyclicpressure pressure in order in order to stopto stop and and examine examine it periodically.it periodically. The The scheduled scheduled loading loading scheme scheme isis shownshown in in Figure4 4.. AA shortershorter load cycle was done done before before formal formal test. test. The The maximum maximum pressure pressure set set for for this this cycle cycle waswas 75 75MPa. MPa. The The ultimate ultimate goal goal of ofthe the pressure pressure test test was was to tocheck check if the if the tank tank could could endure endure the the external external pressurepressure of of126.5 126.5 MPa MPa (i.e., (i.e., 115 115 MPa MPa for for the the 11 11,000,000 m mdeep deep-sea-sea environment environment times times a safety a safety factor factor of of 1.11.1 for for pressure pressure testing). testing). After After test test preparation, preparation, the the tank tank with with strain strain gauges gauges pasted pasted outside outside the the shell shell in inthree three typical typical positions positions was was sent sent to tothe the pressure pressure chamber, chamber, then then the the first first cycle cycle for for watertight watertight and and insulationinsulation tests tests combining combining with with strain strain and and stress stress measurements measurements was was conducted. conducted. The The loading loading phase phase waswas carried carried out out gradually gradually with with the the step step of of pressure pressure increase increase of 5 MPa andand maintainedmaintained forfor 3 3 min min in in the theloading loading and and unloading unloading stage. stage. It wasIt was presumed presumed that that when when the the first first cycle cycle was was finished, finished, the tank the shouldtank shouldbe taken be taken out from out thefrom pressure the pressure chamber chamber for a further for a further inspection. inspection. If there If was there no was problem, no problem, the tank the was tanksent was to thesent chamber to the chamber for a final for test.a final test.

Cycle 2:126.5MPa

Cycle 1:strain&stress P measure

5MPa

3min Time

Figure 4. Loading scheme of pressure test. Figure 4. Loading scheme of pressure test. However, when the tank was taken from the chamber after the first load cycle, it was found that theHowever, flat cover when is cracked the tank as shown was taken in Figure from5 the. Local chamber fracture after occurred the first atload the cycle, top part it was of the found flat that cover. theThe flat upper cover surfaceis cracked and as fracture shown surface in Figure of the5. Local flat cover fracture are respectivelyoccurred at the shown top inpart Figure of the5a,b. flat Based cover. on Thethe upper macroscopic surface and observation, fracture surface it was found of the thatflat cover the flat are cover respectively bore large shown stress in from Figure outside 5a,b. toBased inside, onand the stressmacroscopic concentration observation occurred, it was at the found right-angle that the transition flat cover root. bore The large crack stress morphology from outside presented to inside,intergranular and stress cracking, concentration which wasoccur ar brittleed at fracture.the right-angle transition root. The crack morphology presentFractureed intergranular occurred cracking, before the which maximum was a pressure brittle fracture. reached 75 MPa; however, it is difficult to assess at what pressure level the failure starts. The strain–load curves in Figure6 show some abnormal phenomenon. Position 1 is at the opening reinforcement part of the spherical hull; position 2 is above the equatorial flange of the spherical hull; position 3 is at the bottom of the spherical hull. At a pressure of 40 MPa, there is a relatively obvious change of strain rate in the curves for three recording positions. This indicates that under this loading condition, the whole stress state of the spherical shell changes. At this point, the obvious stress concentration may appear on the flat cover. Because the external load is distributed both on the outer surface of the spherical shell and the flat cover, the change of stress state of the flat cover will cause stress concentration on the shell. From the fracture surface position, the stress concentration may come from the existence of crack-like defects, such as inadequate chamfer and scratches when the sharp part was processed. Therefore, in the next section, defects and fracture analysis will be carried out by means of microscopic observation and finite element calculation.

Appl. Sci. 2019, 9, x FOR PEER REVIEW 6 of 17

47.5 mm

40 mm

(a) (b) Appl. Sci. 2019, 9, 5258 6 of 15 Appl. Sci. 2019, 9, x FOR PEER REVIEW 6 of 17 Figure 5. Photography of the cracked flat cover: (a) upper surface of the cover; (b) fracture surface in the outer edge of the cover.

Fracture occurred before the maximum pressure reached 75 MPa; however, it is difficult to assess at what pressure level the failure starts. The strain–load curves in Figure 6 show some abnormal phenomenon. Position 1 is at the opening reinforcement part of the spherical hull; position 2 is above the equatorial flange of the spherical hull; position 3 is at the bottom of the spherical hull. At a pressure of 40 MPa, there is a relatively obvious change of strain rate in the curves for three recording positions. This indicates that under this loading condition, the whole stress state of the spherical shell changes. At this point, the obvious stress concentration may appear on the flat cover. Because the external load is distributed both on the outer surface of the spherical shell and the flat cover, the

change of stress state of the flat cover will cause stress concentration on the47. 5shell. mm From the fracture surface position, the stress concentration may come from the existence of crack-like defects, such as inadequate40 mm chamfer and scratches when the sharp part was processed. Therefore, in the next section, defects and fracture analysis will be carried out by means of microscopic(b) observation and finite element calculation.(a) Figure 5. Photography of the cracked ﬂat cover: (a) upper surface of the cover; (b) fracture surface in Figure 5. Photography of the cracked flat cover: (a) upper surface of the cover; (b) fracture surface in the outer edge of the cover. the outer edge of the cover. 0.0045 Fracture occurred before the maximum pressure reached 75 MPa; however, it is difficult to assess 0.0040 Position 1 at what pressure level the failure starts. The strain–load curves in Figure 6 show some abnormal Position 2 phenomenon. Position 1 is0.0035 at the opening reinforcement part of the spherical hull; position 2 is above Position 3 the equatorial flange of the spherical hull; position 3 is at the bottom of the spherical hull. At a 0.0030 pressure of 40 MPa, there is a relatively obvious change of strain rate in the curves for three recording positions. This indicates that0.0025 under this loading condition, the whole stress state of the spherical shell changes. At this point, the obvious stress concentration may appear on the flat cover. Because the 0.0020 external load is distributedStrain both on the outer surface of the spherical shell and the flat cover, the change of stress state of the0.0015 flat cover will cause stress concentration on the shell. From the fracture surface position, the stress concentration may come from the existence of crack-like defects, such as 0.0010 inadequate chamfer and scratches when the sharp part was processed. Therefore, in the next section, defects and fracture analysis0.0005 will be carried out by means of microscopic observation and finite element calculation. 0.0000 10 40 55 70 60 30 0 Pressure (MPa)

Figure 6. Recorded0.0045 strains in the three positions outside the shell with variations of pressure.

3. Fracture Examination0.0040 of the FlatPosition Cover 1 Position 2 0.0035 Before the tank was manufactured,Position the3 material from the same furnace was tested to verify its conformance to the specified0.0030 material standard and also the requirements of the designer. However, the same furnace material also had sampling uncertainty. Then in this section the tensile properties 0.0025 of the actual materials used in the flat cover were re-examined by sampling from the broken parts. 0.0020 The non-metal inclusionsStrain analysis, micro-structure analysis, and fracture surface analysis were conducted to find out0.0015 the possible cracking cause.

3.1. Re-Examination of0.0010 Tensile Properties The static tensile0.0005 test of a standard specimen with a diameter of 5 mm and gauge length of 30 mm [16] was carried0.0000 out on SANS HST5106 hydraulic universal testing machine. Table3 shows the results of the tests. It was found10 that40 the55 material70 has60 a higher30 tensile0 strength, yield strength, elongation, and reduction of area comparedPressure to the material (MPa) presented in Table2.

3.2. Re-Examination of Impact Properties

Standard Charpy V-notched impact specimens with length of 55 mm and square cross-sections of 10 mm 10 mm [17] were used to determine impact toughness at room temperature on the PIT-752H × Appl. Sci. 2019, 9, 5258 7 of 15 test machine. Table4 presents the room temperature impact test results of the flat cover material. The fluctuation between the test values is small, and the average impact energy is 14 J. However, the material has low impact toughness. The minimum value of impact toughness of 18Ni(350) is 12 J while the impact toughness of 18Ni(250) is about 24 J, which may be one of the reasons for the cracking of the flat cover during the test.

Table 3. Re-examination results of tensile properties of ﬂat cover material at room temperature.

Specimen σy (MPa) σb (MPa) A (%) Z (%) E (MPa) υ L1 2341 2421 6.4 47 186 0.300 L2 2355 2423 6.6 42 188 0.310 Average 2348 2423 6.5 45 187 0.305

Table 4. Re-examination results of impact toughness of ﬂat cover material at room temperature.

Test Point 1 2 3 Average KV, J 12 16 14 14

3.3. Re-Examination of Hardness Rockwell and Brinell Hardness (HRC) hardness test was carried out with Rockwell hardness tester. The load was 150 KGF with holding time of 10 s. The average hardness of the ﬂat cover material was about 54 HRC.

3.4. Non-Metallic Inclusions Observation Non-metallic inclusions in flat cover samples were inspected [18]. The results of non-metallic inclusions rating and the inclusion morphology in the most important part of the flat cover showed that Appl. Sci. 2019, 9, x FOR PEER REVIEW 8 of 17 there were fewer kinds of non-metallic inclusions. The inclusions specified in the standards include sulfidestate inclusions, of the material alumina meets inclusions, the design requirements. silicate inclusions, Figure and7 shows spherical the inclusion oxides. morphology The original of the state of the materialsample. meets the design requirements. Figure7 shows the inclusion morphology of the sample.

Figure 7. Figure 7.Inclusion Inclusion morphologymorphology of of the the sample sample.. 3.5. Microstructure Observation 3.5. Microstructure Observation MetallographicMetallographic specimens specimens were were taken taken fromfrom the sealing cover. cover. After After mechanical mechanical grinding grinding and and polishingpolishing along along the the longitudinal longitudinal section, section, the the specimens specimens were were etched etched by by an an alcohol alcohol solution solution of FeCl of FeCl3 + 3 + hydrochlorichydrochloric acid acid at room at room temperature. temperature. The The structurestructure was was observed observed under under a ZEISS a ZEISS Axiovert Axiovert 200MAT 200MAT opticaloptical microscope microscope (OM). (OM). The The grading grading method method ofof the grain grain size size of of samples samples was was carried carried out according out according to grainto grain size size rating rating determination determination standard standard [ 19[19].]. FigureFigure 8 shows the the structure structure of ofthe the flat ﬂat cover cover material material whichwhich is martensite. is martensite.

(a) (b)

Figure 8. Metallographic structure of the cover sample: (a) low magnification; (b) high magnification.

Microstructure and morphology of the flat cover material were observed and analyzed by TESCAN VEGA3 XMU scanning electron microscopy (SEM). Figure 9 shows the SEM low and high magnification microstructures of the specimens. The martensite structure of the material was uniform, and a small amount of precipitated phase particles could be seen at the grain boundary.

Appl. Sci. 2019, 9, x FOR PEER REVIEW 8 of 17

state of the material meets the design requirements. Figure 7 shows the inclusion morphology of the sample.

Figure 7. Inclusion morphology of the sample.

3.5. Microstructure Observation Metallographic specimens were taken from the sealing cover. After mechanical grinding and polishing along the longitudinal section, the specimens were etched by an alcohol solution of FeCl3 + hydrochloric acid at room temperature. The structure was observed under a ZEISS Axiovert 200MAT optical microscope (OM). The grading method of the grain size of samples was carried out according Appl.to Sci. grain2019 ,size9, 5258 rating determination standard [19]. Figure 8 shows the structure of the flat cover material8 of 15 which is martensite.

(a) (b)

Figure 8. Metallographic structure of the cover sample: (a) low magnification; (b) high magnification. Figure 8. Metallographic structure of the cover sample: (a) low magnification; (b) high magnification. Microstructure and morphology of the flat cover material were observed and analyzed by TESCAN Microstructure and morphology of the flat cover material were observed and analyzed by VEGA3 XMU scanning electron microscopy (SEM). Figure9 shows the SEM low and high magnification TESCAN VEGA3 XMU scanning electron microscopy (SEM). Figure 9 shows the SEM low and high microstructures of the specimens. The martensite structure of the material was uniform, and a small magnification microstructures of the specimens. The martensite structure of the material was amountAppl.uniform, Sci. of precipitated2019 and, 9, ax FORsmall PEER phaseamount REVIEW particles of precipitated could bephase seen particles at the graincould boundary.be seen at the grain boundary.9 of 17

(b) (a)

Figure 9. SEM microstructure of the cover sample: (a) low magnification; (b) high magnification. Figure 9. SEM microstructure of the cover sample: (a) low magnification; (b) high magnification. Further observation was conducted under JEOL JSM-2100 transmission electron microscopy (TEM). Further observation was conducted under JEOL JSM-2100 transmission electron microscopy Figure 10 shows the TEM morphology of the specimen. It can be seen that the width of martensite lath (TEM). Figure 10 shows the TEM morphology of the specimen. It can be seen that the width of µ is aboutmartensite 0.2–0.8 lathm, is about and a0.2 large–0.8 numberμm, and a of large precipitations number of precipitations were uniformly were distributed uniformly distributed in martensite matrixin martensite with high matrix density. with Electron high density diff.raction Electron analysis diffraction results analysis show results that show a large that number a large number of rod-like ’ particlesof rod are-like Ni particles3(Al, Ti)-type are Ni3(Al,γ precipitates, Ti)-type γ’ precipitates, as shown as in shown Figure in 10 Figureb,c. In 10 addition,b,c. In addition, large sphericallarge granulesspherical in Figure granules 10c in are Fi Fegure2Mo 10 precipitates.c are Fe2Mo Some precipitates. studies Some [20] have studies shown [20] thathave the shown precipitations that the of Fe2Moprecipitation occur withs of the Fe2 increaseMo occur ofwith aging the timeincrease and of aging aging temperature.time and aging temperature.

3.6. Fracture Surface Observation The fracture surface morphology was observed and analyzed by scanning electron microscopy after ultrasonic cleaning of the sample. Figure 11 shows the fracture morphology in the vicinity of the fracture source. It can be seen at the low magniﬁcation (Figure 11a,b) that the crack source is located at the root of the transition between the upper cover and the edge liner. It can be seen at high magniﬁcation (Figure 11c,d) that the crack source has an intergranular morphology and shows brittle

(a) (b) (c)

Figure 10. TEM microstructure of sealed top cover sample: (a) low magnification; (b) high magnification; (c) precipitated phase Fe2Mo morphology.

3.6. Fracture Surface Observation The fracture surface morphology was observed and analyzed by scanning electron microscopy after ultrasonic cleaning of the sample. Figure 11 shows the fracture morphology in the vicinity of the fracture source. It can be seen at the low magnification (Figure 11a,b) that the crack source is located at the root of the transition between the upper cover and the edge liner. It can be seen at high magnification (Figure 11c,d) that the crack source has an intergranular morphology and shows brittle fracture. During the external pressure test, the flat cover was subjected to a large stress from the

Appl. Sci. 2019, 9, x FOR PEER REVIEW 9 of 17

(b) (a)

Figure 9. SEM microstructure of the cover sample: (a) low magnification; (b) high magnification.

Further observation was conducted under JEOL JSM-2100 transmission electron microscopy Appl.(TEM). Sci. 2019 Figure, 9, 5258 10 shows the TEM morphology of the specimen. It can be seen that the width 9 of of 15 martensite lath is about 0.2–0.8 μm, and a large number of precipitations were uniformly distributed in martensite matrix with high density. Electron diffraction analysis results show that a large number fracture. During the external pressure test, the ﬂat cover was subjected to a large stress from the outside of rod-like particles are Ni3(Al, Ti)-type γ’ precipitates, as shown in Figure 10b,c. In addition, large to the inside, and a large stress concentration was generated at the right-angle transition root in which spherical granules in Figure 10c are Fe2Mo precipitates. Some studies [20] have shown that the a crack source was easily generated and resulted in fracture. precipitations of Fe2Mo occur with the increase of aging time and aging temperature.

Appl. Sci. 2019, 9, x FOR PEER REVIEW 10 of 17

outside to the inside, and a large stress concentration was generated at the right-angle transition root in which a crack source was easily generated and resulted in fracture. (a) (b) (c)

FigureFigure 10. 1TEM2 shows microstructure the SEM oftopography sealed top coverof the sample: tearing (a area) low of magniﬁcation; the flat cover. (b) highThe magniﬁcation;tearing surface has aFigure dimple 10. morphology, TEM microstructure which shows of sealed ductile top fracture cover scharacteristics.ample: (a) low magnification; (b) high (c) precipitated phase Fe2Mo morphology. magnification; (c) precipitated phase Fe2Mo morphology.

3.6. Fracture Surface Observation The fracture surface morphology was observed and analyzed by scanning electron microscopy after ultrasonic cleaning of the sample. Figure 11 shows the fracture morphology in the vicinity of the fracture source. It can be seen at the low magnification (Figure 11a,b) that the crack source is located at the root of the transition between the upper cover and the edge liner. It can be seen at high magnification (Figure 11c,d) that the crack source has an intergranular morphology and shows brittle fracture. During the external pressure test, the flat cover was subjected to a large stress from the

(b) (a)

Figure 11. SEM morphology of crack source at the fracture surface of the flat cover: (a) low magnification Figure 11. SEM morphology of crack source at the fracture surface of the flat cover: (a) low morphology; (b) low magnification morphology at the crack source; (c) high magnification morphology magnification morphology; (b) low magnification morphology at the crack source; (c) high at the crack source; (d) partially magnified morphology at the crack source. magnification morphology at the crack source; (d) partially magnified morphology at the crack source.

Appl. Sci. 2019, 9, 5258 10 of 15

Appl. Sci.Figure 2019 ,12 9, xshows FOR PEER the REVIEW SEM topography of the tearing area of the ﬂat cover. The tearing surface11 of has 17 aAppl. dimple Sci. 2019 morphology,, 9, x FOR PEER which REVIEW shows ductile fracture characteristics. 11 of 17

(b) (a) (b) (a) Figure 12. Topography of cracking area of the flat cover: (a) low magnification morphology; (b) high Figure 12. Topography of cracking area of the flat cover: (a) low magnification morphology; (b) high magnification morphology. magnificationFigure 12. Topography morphology. of cracking area of the flat cover: (a) low magnification morphology; (b) high magnification morphology. 4.4. Finite Element Analysis 4. Finite Element Analysis Based on the observations in Section3 3,, itit cancan bebe concludedconcluded thatthat thethe finalfinal fracturefracture ofof thethe flatflat covercover is due dueBased to to the the on large largethe stressobservations stress concentration concentration in Section generated generated3, it can at be the atconcluded right-angle the right that-angle transition the transitionfinal root fracture of root the of thicker of the the flat thicke central coverr partiscentral due and topart the the thinnerand large the stressedge. thinner A concentrationcrack edge. initiated A crack generated from initiated this at point from the is rightthis shown -pointangle in Figureis transition shown 13. in It root wasFigure of initiated the 13 . thickeIt fromwasr impropercentralinitiated part from manufacture and improper the thinner and manufacture has edge. the length A crackand of has 3initiated mm the and length from the depth ofthis 3 mmpoint of 1.5 and is mm. shownthe Final depth in failure Figureof 1.5 easily mm13. occurs.It F wasinal becauseinitiatedfailure easily thefrom material occursimproper hasbecause manufacture low impact the material and and toughness has thelow length impact properties. of and 3 mm toughness Apresumed and the properties.depth semi-elliptical of 1.5 A mm presumed surface. Final crackfailuresemi-elliptical is easily considered occurs surface in because thiscrack section is theconsidered to material analyze in has thethis low failure section impact mode to analyze and of thetoughness flatthe failure cover properties. during mode loadingof the A presumedflat process. cover Thesemiduring fracture-elliptical loading criterion surface process. is crack thatThe theisfracture considered stress criterion intensity in this is factor that section the at the stressto analyze crack intensity tip the reaches failurefactor the atmode the plane crack ofstrain the tip flat fracturereaches cover toughnessduringthe plan loadinge strain of the process. fracture material. Thetoughness fracture of criterion the material is that. the stress intensity factor at the crack tip reaches the plane strain fracture toughness of the material.

Figure 13. Crack position at the right-angle transition root. Figure 13. Crack position at the right-angle transition root. 4.1. Fracture Toughness of the Material Figure 13. Crack position at the right-angle transition root. 4.1. Fracture Toughness of the Material Fracture toughness tests are not conducted in the present study; however, it can be deduced from4.1. FractureF theracture widely Toughness toughness accepted of thetests assumptions Material are not conducted that the fracture in the toughness present study; is taken however, to be a criticalit can be value deduced of the maximumfromF theracture widely stress toughness intensityaccepted tests factor assumptions are for not unstable conducted that crackthe fracture growth in the toughness [present21]. According study; is taken however, to thisto be definition, a criticalit can be thevalue deduced unstable of the K K crackfrommaximum the growth widely stress will accepted intensityoccur when assumptions factormax forreaches unstable that theIC .fracture crack In order growth toughness to reduce [21]. istest According taken costs, to thisbe to a method critical this definition, wasvalue used of the to determinemaximumunstable crack the stress fracturegrowth intensity toughnesswill occur factor ofwhen for the unstable materialKmax reaches crack conveniently K IC growth. In orde on [2r1 theto]. reduce According basis of test the costs, to crack this this growth definition, method rate wa test thes forused data to determine over a wide the range fracture of growth toughness rates of [ 9the] at material four load conveniently ratios, i.e., 0.1, on 0.3,the basis 0.5, and of the 0.7. crack growth unstable crack growth will occur when Kmax reaches KIC. In order to reduce test costs, this method was usedrate test to determine for data over the fracturea wide range toughness of growth of the rates material [9] at conveniently four load ratios, on the i.e., basis 0.1, of 0.3, the 0.5, crack and growth 0.7. rate test for data over a wide range of growth rates [9] at four load ratios, i.e., 0.1, 0.3, 0.5, and 0.7.

Appl. Sci. 2019, 9, x FOR PEER REVIEW 12 of 17

Appl. Sci. 2019, 9, 5258 11 of 15 The crack growth rate curves (da/dN~ΔK curves) at four load ratios of 0.1, 0.3, 0.5, and 0.7 were experimentally obtained in [9]. At each load ratio, two groups of tests were carried out. When the crackThe growth crack growth curve rate began curves to transit (da/dN ~ from∆K curves) stable at growth four load stage ratios to ofunstable 0.1, 0.3, growth 0.5, and stage, 0.7 were the experimentallyrequirement ofobtained Kmax = KICin should [9]. At be each satisfied. load Then ratio, a two value groups of KIC of can tests be calculated were carried based out. on When each crack the crackgrowth growth rate curvecurve. began The average to transit of from these stable values growth can stagebe taken to unstable as the growthfinal value stage, of theKICrequirement as shown in ofFigureKmax =14.K ICThshoulde data of be Δ satisﬁed.K(1 − R) Thenare within a value a relatively of KIC can stable be calculated range, and based the on fracture each crack toughness growth of 1/2 rate18Ni curve. (350)The can averageapproach of 37~4 these0 MPa· valuesm can, which be taken agrees as the well ﬁnal with value the data of K ICrangeas shown in literature in Figure [12,13] 14. Theand data can ofbe∆ appliedK(1 R )for are fracture within analysis. a relatively The stable dotted range, line andshows the the fracture average toughness level of ofthe 18Ni calculated (350) − candata approach of KIC, which 37~40 was MPa aboutm1/2, 38 which MPa· agreesm1/2. It well must with be emphasized the data range that in the literature toughness [12, 13of] this and mat canerial be · appliedis greatly for reduced fracture compared analysis. Theto 18Ni dotted (250 line), which shows is thetraditionally average level used of inthe the calculatedpressure hulls data of of deepKIC, - whichsea manned was about submersible 38 MPa m [9].1/2 . It must be emphasized that the toughness of this material is greatly · reduced compared to 18Ni (250), which is traditionally used in the pressure hulls of deep-sea manned submersible [9].

) 40 0.5

35 KIC

) (MPa m (MPa ) 30 R

(1- 25

 , ,

K 20 

15 K K (1-R) 10 0.1 0.3 0.5 0.7 R

Figure 14. ∆K and ∆K(1 R) values of 18Ni(350) determined by unstable crack growth properties. − 4.2. FiniteFigure Element 14. Δ AnalysisK and ΔK(1 − R) values of 18Ni(350) determined by unstable crack growth properties.

4.2. SolidWorksFinite Element software Analysis was used for three-dimensional modeling, and ABAQUS software is used for finite element analysis after modeling. C3D8R element is used for meshing. Local mesh encryption was madeSolidWorks in the vicinity software of keywaschamfers used for three of the-dimensional flat cover and modeling, contacts theand part ABAQUS between software the spherical is used modelfor finite and element the flat cover,analysis with after mesh modeling. size of C3 1 mm,D8R comparedelement is toused the for mesh meshing size of. Local 20 mm mesh in other encryption parts. Thewas small made chamfer in the vicinity was not of included key chamfers in the of calculation the flat cover model, and contacts because whenthe part the between crack was the insertedspherical intomodel the and model, the theflat stresscover, concentration with mesh size at of the 1 mm, crack compared tip would to be the the mesh dominant size of factor,20 mm and in other the stress parts. concentrationThe small chamfer at the was small not chamfer included was in nothe longer calculation important. model, The because whole when structure the crack model was was inserted used ininto finite the element model, calculationthe stress concentration to obtain the at critical the crack stress tip area would and be the the surface dominant crack factor, was insertedand the stress into theconcentration local model at after the stresssmall analysis.chamfer was Figure no longer15 gives important. the whole The model whole for structure stress analysis. model Thewas ballastused in tankfinite used elemen in thet calculation deep-sea submersible to obtain the was critical subjected stress area to uniformly and the surface distributed crack hydrostatic was inserted pressure, into the whichlocal model was applied after stress both analysis. in the spherical Figure 1 shell5 gives and the the whole flat cover.model for In orderstress toanalysis. eliminate The theballast overall tank rigidused body in the displacement deep-sea submersible without hindering was subjected the relative to uniforml deformation,y distributed the boundaryhydrostatic conditions pressure, werewhich setwas symmetrically applied both with in the six spherical displacement shell and components the flat cover. constrained In order by to the eliminate three-point the overall support rigid method body recommendeddisplacement in without [22]. The hindering contact surface the relative between deformation, the flat cover and the the boundary sphere was conditions set as the were friction set contactsymmetrically with the withfriction six coe displacementfficient of 0.1. components constrained by the three-point support method recommended in [22]. The contact surface between the flat cover and the sphere was set as the friction contact with the friction coefficient of 0.1.

Appl. Sci. 2019, 9, x FOR PEER REVIEW 13 of 17 Appl. Sci. 2019, 9, 5258 12 of 15 Appl. Sci. 2019, 9, x FOR PEER REVIEW 13 of 17 Friction constraints： μ =0.1 Friction constraints： μ =0.1

UY=UZ=0 UY=UZ=0

Y Z X

Z X UX=UZ=0 UX=UZ=0 Figure 15. Finite element model for stress analysis. Figure 15. Finite element model for stress analysis. Figure 15. Finite element model for stress analysis. When a crack is simulated, the crack will be inserted into a sub-model. The sub-model was taken When a crack is simulated, the crack will be inserted into a sub-model. The sub-model was taken out separately for the study. The calculation can be conducted by using the connection between the out separatelyWhen a crack for the is simulated, study. The the calculation crack will can be inserted be conducted into a bysub using-model. the The connection sub-model between was taken the global model and the sub-model, thus greatly reducing unnecessary computation and time costs. globalout separately model and for the study. sub-model, The calculation thus greatly can reducing be conducted unnecessary by using computation the connection and between time costs. the According to the results of previous failure observations, the defective part of the cover is treated as Accordingglobal model to the and results the sub of- previousmodel, thus failure greatly observations, reducing theunnecessary defective partcomputation of the cover and is time treated costs. as a sub-model and the crack with the length of 3 mm and the depth of 1.5 mm is inserted into the model. aAccording sub-model to and the the results crack of with previous the length failure of observations, 3 mm and the the depth defective of 1.5 mm part is of inserted the cover into is thetreated model. as After static analysis, the stress intensity factor of crack front can be obtained. The position relationship Aftera sub static-model analysis, and the the crack stress with intensity the length factor of 3 of mm crack and front the depth can be of obtained. 1.5 mm is The inserted position into relationship the model. between the sub-model and the global model is shown in Figure 16. betweenAfter static the analysis, sub-model the andstress the intensity global model factor isof showncrack front in Figure can be 16 obtained.. The position relationship between the sub-model and the global model is shown in Figure 16.

Figure 16. The position relationship between the sub-model and the global model. Figure 16. The position relationship between the sub-model and the global model.

When theFigure ﬂat cover 16. The is subjected position relationship to diﬀerent between external the pressures sub-model from and 5 tothe 85 global MPa, model. the stress intensity factors of the crack can be obtained. Figure 17 depicts the stress intensity factors along the crack front with increasing the external pressure. Figure 18 shows the changing of the stress intensity factors at

Appl. Sci. 2019, 9, x FOR PEER REVIEW 14 of 17

When the flat cover is subjected to different external pressures from 5 to 85 MPa, the stress intensity factors of the crack can be obtained. Figure 17 depicts the stress intensity factors along the Appl.crack Sci. front 2019 with, 9, x FOR increasing PEER REVIEW the external pressure. Figure 18 shows the changing of the stress intensity14 of 17 factors at the surface points and the deepest point with the external pressure. The stress intensity factorWhen in the the length flat coverdirection is subjected is larger tothan different that in externalthe depth pressures direction. from The 5maximum to 85 MPa, value the ofstress the intensitystress intensity factor sfactor of the occurs crack canat the be surfaceobtained. points. Figure When 17 depict the externals the stress load intensity increases factors to about along 75 –the85 crackMPa, frontthe stress with increasingintensity fa thector external at the crackpressure. tip Figurereaches 1 835 shows–42 MPa· the mchanging1/2, i.e., approximating of the stress intensity to the factors at the surface points and the deepest1/2 point with the external pressure. The stress intensity Appl.valu Sci.e 2019 of fracture, 9, 5258 toughness, 38 MPa·m , shown in Figure 14, which is the ultimate cause13 of of 15 the factordamage. in the length direction is larger than that in the depth direction. The maximum value of the stress intensity factor occurs at the surface points. When the external load increases to about 75–85 theMPa, surface the pointsstress intensity and the deepest factor at point the withcrack the tip external reaches pressure.35–42 MPa· Them stress1/2, i.e., intensity approximating factor in to the the lengthvalue direction of fracture is larger toughness than, that 38 inMPa· them depth1/2, shown direction. in Figure The maximum 14, which value is the of ultimate the stress cause intensity of the factordamage. occurs at the surface45 points. When the external load increases to about 75–85 MPa, the stress intensity factor at the crack tip reaches 35–42 MPa m1/2, i.e., approximating to the value of fracture KIC · toughness, 38 MPa m1/240, shown in Figure 14, which is 5MPa the ultimate cause 15MPa of the damage. · crack 35 25MPa 35MPa

) 45 45MPa 55MPa 0.5 30 K 65MPa 75MPa 40 IC crack 5MPa85MPa 15MPa 25 35 25MPa 35MPa

) 45MPa 55MPa

(MPa m (MPa 20 0.5 K 30 65MPa 75MPa 15 85MPa 25 10 (MPa m (MPa 20 K 5 15 0 100.0 0.2 0.4 0.6 0.8 1.0 5 Normalized distance from one of the crack tips

0 0.0 0.2 0.4 0.6 0.8 1.0 Figure 17. Stress intensity factor along crack front with increasing external pressure. Normalized distance from one of the crack tips

Figure 17. Stress intensity factor along crack front with increasing external pressure.

Figure 17. Stress45 intensity factor along crack front with increasing external pressure. K 40 IC 35

) 45

0.5 30 length K 40 IC 25 depth 35

) (MPa m (MPa K

0.5 30 length 15 25 depth 10

20 (MPa m (MPa K 5 15 0 10 0 10 20 30 40 50 60 70 80 90 5 Pressure (MPa)

Figure 18. Stress0 intensity factor at the surface points and the deepest point. 0 10 20 30 40 50 60 70 80 90 5. Summary and Conclusions Pressure (MPa) In this paper, a destroyed flat cover made of 18Ni (350) used in an adjustable ballast tank during high pressure testing was analyzed. To find out the cause of the damage, the mechanical properties of the material used in the flat cover were re-examined by sampling and testing of the broken parts. Non-metallic inclusions analysis, microstructure analysis, and fracture surface analysis were conducted to find out the possible fracture cause. Finite element analysis based on fracture mechanics to understand the ultimate cause of destruction was conducted. The following conclusions and recommendations can be made: Appl. Sci. 2019, 9, 5258 14 of 15

(1) Maraging steels are expected to still have high toughness when improving their ultra-high strength level. 18Ni (250) grade is currently an acceptable candidate material for deep-sea pressure hulls, which has achieved satisfactory application experience. To obtain better weight-to-displacement ratio, the designers applied 18Ni (350) for an adjustable ballast tanks used in a deep-sea submersible. The strength of 18Ni (350) is higher than that of 18Ni (250), however its toughness is lower. In the current design guideline for deep-sea pressure hulls, the requirements for material toughness have not been emphasized. The trial of applying 18Ni (350) was made in the present study but unexpected failure of the flat cover occurred when load was relatively small. (2) Non-metal inclusions analysis, microstructure analysis, and fracture surface analysis showed that the original state of the material meets the design requirements with fewer kinds of non-metallic inclusions. The material has low impact toughness, which is one of the reasons for the cracking of the flat cover during the test. (3) From the design point of view, it is suggested to optimize the transition from right angle to chamfer at variable cross-sections, and to increase the thickness of the outer edge liner of the cover appropriately, which will reduce the possibility of damage. During processing, the surface finish of structural parts should be ensured, especially the root surface of the variable cross-section, so as to avoid surface processing defects. (4) The reliability of deep-sea pressure hulls, especially the pressure hulls of manned submersibles, guarantees the safety of personnel and equipment. There must be attention paid to the comprehensive performance of materials. From the point of view of material selection, the application of 18Ni (350) maraging steel to the pressured structure in deep-sea environments is a risky choice. On the other hand, the 18Ni series of maraging steels have broad application prospects in deep-sea environments, so improving the toughness and plasticity together with increasing the strength level will be an important research direction [23].

Author Contributions: The individual contributions are specified as follows: conceptualization, F.W. and W.C.; methodology, F.W. and J.Z.; software, M.W.; validation, M.W., S.Z., G.T., and Z.J.; formal analysis, F.W.; investigation, S.Z.; resources, Z.J.; data curation, M.W. and G.T.; writing—original draft preparation, F.W.; writing—review and editing, W.C.; visualization, F.W.; supervision, W.C.; project administration, F.W. and W.C.; funding acquisition, F.W. and W.C. Funding: This research was funded by the General Program of National Natural Science Foundation of China, grant number 51679133, the State Key Program of National Natural Science of China, grant number 51439004, and the “Construction of a Leading Innovation Team” project by the Hangzhou Municipal government. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Nanba, N.; Morihana, H.; Nakamura, E.; Watanabe, N. Development of Deep Sub-mergence Research Vehicle “Shinkai 6500”. Tech. Rev. 1990, 27, 157–168. 2. Walden, B.B.; Brown, R.S. A replacement for the Alvin submersible. Mar. Technol. Soc. J. 2004, 38, 85–91. [CrossRef] 3. Liu, Y.S.; Wu, D.F.; Li, D.L.; Zhao, X.F. Seawater hydraulic buoyancy adjusting system for large-depth submersible. Chin. Hydraul. Pneum. 2014, 10, 1–10. 4. Qiu, Z.L. Design and research on a variable ballast system for deep-sea manned submersibles. J. Mar. Sci. Appl. 2008, 7, 255–260. [CrossRef] 5. Tangirala, S.; Dzielski, J. A variable buoyancy control system for a large AUV. IEEE J. Ocean. Eng. 2007, 32, 762–771. [CrossRef] 6. Jame, J.G.; Lisa, L.L. Submersible pressure hull design parametrics. SNAME Trans. 1991, 99, 119–146. 7. Ross, C.T.F. A conceptual design of an underwater vehicle. Ocean Eng. 2006, 33, 2087–2104. [CrossRef] 8. Qu, W.X. Structure Design and Analysis of Pressure Hull in Manned Submersible. Ph.D. Thesis, Harbin Engineering University, Harbin, China, 2013. 9. Wang, F.; Hu, Y.; Cui, W.C. Preliminary evaluation of maraging steels on its application to full ocean depth manned cabin. J. Ship Mech. 2016, 20, 1557–1572. Appl. Sci. 2019, 9, 5258 15 of 15

10. 18 Percent Nickel Maraging Steels-Engineering Properties, Distributed by the Nickel Development Institute, Courtesy to Inco Europe Limited; NiDI Publication: Groningen, The Netherlands, 1976; No.4419. 11. Sagalevitch, A.M. From the Bathyscaph Trieste to the Submersibles Mir. Mar. Technol. Soc. J. 2009, 43, 79–86. [CrossRef] 12. Fan, S.J.; Dai, T. Properties and applications of 18Ni grade maraging steels. Foreign Met. Heat Treat. 1995, 3, 41–47. 13. Gangloff, R.P. Hydrogen assisted cracking of high strength alloys. In Comprehensive Structural Integrity Vol. 6; Milne, I., Ritchie, R.O., Karihaloo, B., Eds.; Elsevier Science: New York, NY, USA, 2003. 14. He, Y.; Liu, K.; Yang, K. Effect of solution temperature on fracture toughness and microstructure of ultra-purified 18ni(350) maraging steel. Acta Met. Sin. 2003, 39, 381–386. 15. ISO 16528-1-2007. Boilers and Pressure Vessels—Part 1: Performance Requirements; ISO: Geneva, Switzerland, 2007. 16. ISO 6892-1-2016. Metallic Materials-Tensile Testing—Part 1: Method of Test at Room Temperature; ISO: Geneva, Switzerland, 2016. 17. ISO 14556-2000. Steel-Charpy V-notch Pendulum Impact Test-Instrumented Test Method; ISO: Geneva, Switzerland, 2000. 18. ISO4967-1998. Steel-determination of Content of Nonmetallic Inclusions-micrographic Method Using Standards Diagrams; ISO: Geneva, Switzerland, 1998. 19. GB/T 6394-2017. Determination of Estimating the Average Grain Size of Metal; China National Standards: Beijing, China, 2017. 20. Mu, B. Effect of Trace Elements on the Mechanical Properties and Microstructure of 18 wt.%Nickel Maraging Steel. Master’s Thesis, Jiangsu University of Science and Technology, Zhenjiang, China, 2011. 21. Farahmand, B.; Nikbin, K. Predicting fracture and fatigue crack growth properties using tensile properties. Eng. Fract. Mech. 2008, 75, 2144–2155. [CrossRef] 22. China Classification Society (CCS). Rules for the Classification and Construction of Diving Systems and Submersibles; China Classification Society: Beijing, China, 2013. 23. Marcisz, J.; Adamczyk, M.; Garbarz, B. Optimisation of Mechanical Properties of 18%Ni350 Grade Maraging Steel Using Novel Heat Treatment. Arch. Metall. Mater. 2017, 62, 73–84. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).