Plus - extended fatigue analysis of ship details

Transcription

1 CLASS GUIDELINE DNVGLCG0152 Edition April 2016 Plus extended fatigue analysis of ship details The electronic pdf version of this document, available free of charge from is the officially binding version.

2 FOREWORD DNV GL class guidelines contain methods, technical requirements, principles and acceptance criteria related to classed objects as referred to from the rules. April 2016 Any comments may be sent by to If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of DNV GL, then DNV GL shall pay compensation to such person for his proved direct loss or damage. However, the compensation shall not exceed an amount equal to ten times the fee charged for the service in question, provided that the maximum compensation shall never exceed USD 2 million. In this provision "DNV GL" shall mean, its direct and indirect owners as well as all its affiliates, subsidiaries, directors, officers, employees, agents and any other acting on behalf of DNV GL.

3 Changes current CHANGES CURRENT This is a new document. Class guideline DNVGLCG0152. Edition April 2016 Page 3 Plus extended fatigue analysis of ship details

4 Changes current... 3 Section 1 Introduction Objective General Symbols and abbreviations... 6 Section 2 Scope of class notation Plus Longitudinal stiffenerframe connections Bottom and side shell plating Stringer heel and toe details Deck openings and deck details Ship specific details Low cycle fatigue...12 Section 3 Analysis procedure analysis procedure Fatigue analysis Stress analysis Finite element analysis Corrosion additions...13 Section 4 Stress analysis of stiffenerframe connections General Calculation of stress components Loading conditions...14 Section 5 Finite element modelling of stiffenerframe connections General Seminominal stress model Stress concentration models Loads Stress read out from FE models Screening procedure Section 6 Documentation of Plus analysis FEmodels Stresses Class guideline DNVGLCG0152. Edition April 2016 Page 4 Plus extended fatigue analysis of ship details Contents CONTENTS

5 4 Stress concentration factor analysis Section 7 References General Appendix A Stress concentration factors General Establish stress concentration factors Example of establishing stress concentration factors Stress concentration factors for typical longitudinal end connection details Seminominal finite element mesh of stiffenerframe connections...48 Changes historic Class guideline DNVGLCG0152. Edition April 2016 Page 5 Plus extended fatigue analysis of ship details Contents 3 Fatigue calculations... 30

6 1 Objective This class guideline provides guidance on how to perform and document analyses required for compliance with the class notation Plus as described in the RU SHIP Pt.6 Ch.1 Sec.6. The class notation provides enhanced scope for fatigue strength and ensures that specific critical structural details are adequately designed to meet specified fatigue requirements. 2 General The Plus notation is an optional class notation mainly intended for vessels operating in harsh areas and includes extended scope of fatigue strength verification for hull structural details. This class guideline covers a description of the following: Scope Procedures for Modelling Structural analysis Stress read out Fatigue analysis Stress concentration factors Documentation of the analysis. Calculations documenting compliance with requirements in this section shall be submitted for verification. The fatigue strength evaluation shall be carried out based on the target fatigue life, service area and loads specified by the CSR for CSR ships and RU SHIP Pt.3 for other ship types. The method of fatigue assessment is given in DNVGL CG The following details in the cargo area are to be considered in the fatigue strength assessment in addition to those required for other class notations: Longitudinal stiffenerframe connections located in the bottom, inner bottom, side and inner side including connected web stiffener, cutout and collar plate, Sec.2 Figure 1. Strength deck plating including stress concentrations from openings, scallops, pipe penetrations and attachments Bottom and side shell plating connection to web frames and stiffeners Stringer heels and toes Ship specific details as specified in Sec.2 [5]. The effect of low cycle fatigue is to be considered for details subjected to large stress variations during loading and unloading. Low cycle fatigue analysis shall be performed according to procedure given in DNVGL CG Symbols and abbreviations The symbols following symbols and abbreviations are used in this class guideline. For symbols not defined in this class guideline, reference is made to RU SHIP Pt.3 Ch.1 Sec.4, RU SHIP Pt.3 Ch.9 and DNVGL CG Class guideline DNVGLCG0152. Edition April 2016 Page 6 Plus extended fatigue analysis of ship details Section 1 SECTION 1 INTRODUCTION

7 Section 2 SECTION 2 SCOPE OF CLASS NOTATION PLUS 1 Longitudinal stiffenerframe connections 1.1 General In general it shall be demonstrated that the fatigue analysis for all stiffenerframe connections in the midship, forward, and aftcargo areas are in compliance with the requirements for notation Plus. Compliance can be demonstrated based on direct analysis or by simplified method, as described below, or a combination of those. 1.2 Direct analysis The connections in bottom, inner bottom, side, inner side and hopper tank may be assessed with finite element analysis according to Sec.5. Figure 1 illustrates the typical hotspots at a stiffenerframe connection. The hotspots include locations on the webframe, stiffener lug and the web stiffener. The analyses should cover the entire cargo area of the selected vessel. See Figure 2 for an overview. Connections in double side and bottom at one webframe in the middle of the forward, amidships and afttank may be analysed according to the procedure described in this class guideline. The fatigue calculations should be performed for the midframe of: Amidships cargo area Aft cargo area Forward cargo area. The fatigue capacity of the other frames where fatigue calculation is not performed may be assessed with the screening procedure described in Sec.5 [6]. 1.3 Simplified method LNG membrane and LPG carrier (with independent tanks of type A or type C): Design for cut outs in case where the web stiffener is omitted or not connected to the longitudinal are required to adopt a full tight collar plate for the following members: Side shell below 1.1Tsc Inner hull longitudinal bulkhead below 1.1Tsc Top side tank sloping plate below 1.1Tsc Hopper and inner bottom Bottom. For cut outs in case where the web stiffener is connected to the longitudinal stiffener flange, the collar plate is to be extended to the attached plate. In addition, for LNG carriers with membrane tanks, keyhole type scallop and soft toe is to be provided for web stiffener unless back bracket is fitted. Where the web stiffener is located on side shell below the lowest stringer and hopper plate, the web stiffener is to have soft toe (LNG membrane carrier only). Longitudinal stiffener on side shell and topside tank which supports the upper connecting side frame brackets shall have a full tight collar plate and web stiffener with soft toed, if any. (LPG carrier only). A full tight collar plate is to be provided for the area prone to high shear stress in way of opening either double bottom or double side structures. Class guideline DNVGLCG0152. Edition April 2016 Page 7 Plus extended fatigue analysis of ship details

8 Cut outs where the web stiffener is not connected to the longitudinal full tight collar plate shall be applied the following areas: Side shell below 0.9D Inner hull longitudinal bulkhead below 0.9D Top side tank sloping plate below 0.9D Hopper and inner bottom Bottom. Cut outs in case where the web stiffener is connected to the longitudinal stiffener flange shall have collar plate extended to the attached plate and keyhole type scallop and soft toe is to be provided for web stiffener (unless back bracket is fitted) for the following members: Side shell below 1.1Tsc Inner hull longitudinal bulkhead below 1.1Tsc Top side tank sloping plate below 1.1Tsc Hopper and inner bottom Bottom. Full tight collar plate is to be provided for the area with high shear stress in way of openings in double bottom and double side structures. Full tight collar plate is to be provided for the longitudinal stiffener in way of bracket toe of primary support member. For bottom and inner bottom longitudinals, full tight collar plate is to be provided for stiffeners within 20% 30% of effective shear span. For forward and aft cargo tanks, full tight collar plate or improved cutout is to be provided for the longitudinal stiffeners between 0.9D and the lowest stringer as well as in way of longitudinal stiffeners in the hopper tank. Alternative design verified by direct analysis may be accepted on a case by case basis. Ship types not covered shall be based on special consideration. Application of structural details described above may be specially considered casebycase for designs which is different from typical arrangement. Class guideline DNVGLCG0152. Edition April 2016 Page 8 Plus extended fatigue analysis of ship details Section 2 Oil tankers with plane bulkheads in way of cargo tanks:

9 Section 2 Figure 1 Hot spots at stiffenerframe connections Figure 2 View of ship and location of areas to be analysed Class guideline DNVGLCG0152. Edition April 2016 Page 9 Plus extended fatigue analysis of ship details

10 Section 2 Figure 3 Web frames in cargo area 2 Bottom and side shell plating The fatigue capacity of the bottom and side shell plating between two webframe positions in midship, forward and aftcargo area shall be assessed using the prescriptive fatigue analysis. Both the transverse stress at stiffener midlength and the longitudinal stress at the platetransverse frame intersection should be assessed. It is sufficient to use simplified stress formulas for plate bending due to lateral pressure given in DNVGL CG The plate field should be subjected to internal and external rule pressure loads as given in RU SHIP Pt.3 Ch.4 or according to CSR BC&OT Pt.1 Ch.4, whichever is relevant. The fatigue analysis shall be based on maximum effective stress range from the considered dynamic load cases and stress concentration factors according to procedures of DNVGL CG Stringer heel and toe details Fatigue calculations should be carried out for the stringers in the midship area, forward and aft hold. The heel and toe of the stringers are normally the critical locations to be assessed. This is illustrated in Figure 4. The fatigue calculations should be performed using a fine mesh finite element model with appropriate t t mesh in order to capture the hotspot stress. The procedures for modelling and stress readout from fine mesh finite element models are described in DNVGL CG 0127 and DNVGL CG The analysis is typically performed using a cargo hold model and a local fine mesh model together with the submodelling technique. Class guideline DNVGLCG0152. Edition April 2016 Page 10 Plus extended fatigue analysis of ship details

11 Section 2 Figure 4 Critical locations on stringers 4 Deck openings and deck details Deck openings and deck details in the cargo area should be analysed with respect to fatigue. Normally it is sufficient to use stress concentration factors for deck details as given in DNVGL CG The details that should be included are: Openings Pipe penetrations Attachments Scallops. The fatigue requirements for the deck plating will normally be satisfied provided that the target fatigue life is obtained with a stress concentration factor Kg = 3.0. Stress concentration model should be made if Kg of the target detail is not found in DNVGL CG 0129 App.A. The fatigue calculation is performed using the nominal stress due to hull girder bending together with relevant stress concentration factor to obtain the hotspot stress. The nominal stress level may be established using the prescriptive method given in DNVGL CG 0129 Sec.4. The control of the deck openings and deck details may have direct impact on the hull girder cross section modulus, ref. RU SHIP Pt.3 Ch.5. 5 Ship specific details For different ship types there are characteristic critical areas of attention which are prone to fatigue and these areas should be included in the fatigue analysis for Plus notation. Since these details vary from vessel type to vessel type these ship specific details need to be selected at an early stage for clarification. Details that require fatigue analysis are: Class guideline DNVGLCG0152. Edition April 2016 Page 11 Plus extended fatigue analysis of ship details

12 Section 2 LPG carriers (prismatic Atype tanks): Lower and upper side brackets Dome opening and coaming. LNG membrane carriers: Upper hopper knuckle Longitudinal girders at transverse bulkheads Upper and lower chamfer knuckles Dome opening and coaming. In general very fine mesh finite element analysis is required to verify the fatigue strength of these details. 6 Low cycle fatigue Low cycle fatigue strength of highly stressed locations under repeated cyclic static loads, mainly due to cargo loading and unloading, should be considered as significant yielding can cause cracks at hotspots even though the dynamic stress from wave loading is low. A procedure for calculating the combined damage due to low cycle and high cycle fatigue is described in DNVGL CG 0129 App.H. The low cycle stress range due to loading and unloading is based on the use of a pseudoelastic hot spot stress range derived by use of a plasticity correction factor on the elastic stress range. The method enables use of standard hotspot SNcurve. For compliance with the Plusnotation the following locations need to be verified with respect to low cycle fatigue: Web stiffener on top of inner bottom longitudinal and hopper slope longitudinals when wide frame space is employed. Webframe hotspots at the stiffenerframe connections in areas of high girder shear stress or where web stiffener is not connected to top of longitudinal stiffener. Heel and toe of horizontal stringer of transverse bulkhead for which frequent alternate loading is anticipated. Inner bottom connection to transverse bulkhead for which frequent alternate loading is anticipated. Lower transverse stool connection to inner bottom for a loading condition with one tank empty and the tank on opposite side full. Class guideline DNVGLCG0152. Edition April 2016 Page 12 Plus extended fatigue analysis of ship details

13 1 Fatigue analysis In order to obtain the class notation Plus all structural details described in the scope should comply with a fatigue damage ratio equal or below 1.0 for the specified design fatigue life. The environmental reduction factor, fe, is not to be taken less than the following: fe fe = 1.0 for ships with class notation CSR = 0.8 for other ship types. 2 Stress analysis The long term stress ranges for fatigue calculations of Plus details are to be calculated according to the procedure given in DNVGL CG The hot spot stress range calculation will be based on the maximum stress range from the considered dynamic load cases: σ = maxi ( σfs, i(j)) where σ 2 = Fatique stress range, in N/mm, for load case (i) of loading condition (j). Further details for fatigue stress range calculation are described in DNVGL CG 0129 Sec.6. 3 Finite element analysis Finite element models and analysis for all details in the scope of Plus notation except the stiffenerframe connections shall be according to DNVGL CG 0129 Sec.6. The procedure for modelling of longitudinal stiffenerframe connections is given in Sec.5. 4 Corrosion additions Net scantlings as defined by RU SHIP Pt.3 Ch.3 or CSR, whichever is relevant, shall be used in the finite element model. For vessels with the class notation ESP and CSR the stresses are to be calculated for a thickness equal tn50. For other vessels the stresses will be based on the modelled gross offered thickness, tgr. The scantling approach factor for correction of stress according to net scantlings approach, fc, are to be applied, see DNVGL CG 0129 Sec.3 [2.2]. Class guideline DNVGLCG0152. Edition April 2016 Page 13 Plus extended fatigue analysis of ship details Section 3 SECTION 3 ANALYSIS PROCEDURE ANALYSIS PROCEDURE

14 1 General The stress analysis is to be done in accordance with hot spot stress approach given in DNVGL CG The local loads from sea pressure, ballast pressure and cargo loads will be dominating in way of these connections and the hull girder loads will have little influence if any. Hence the dynamic load cases representing beam sea for midship area will be sufficient. For forward and aft cargo areas both beam sea and oblique sea cases need to be considered. 2 Calculation of stress components The calculation of the stress range is performed by finite element analysis of a seminominal mm mesh size model. Hotspot stress range for fatigue calculations is found by multiplying the seminominal stress with stress concentration factors. Sec.5 describes the procedure for finite element analysis. 3 Loading conditions Fatigue analyses should be carried out for representative loading conditions. The following two loading conditions are normally sufficient for documentation for analysis of oil tankers, container vessels and gas carriers. Normal full load departure condition Normal ballast arrival condition. Other ship types may require other representative loading conditions in addition to these two conditions, see RU SHIP Pt.3 Ch.9 Sec.4 and CSR Pt.1 Ch.4 Sec.8 [5]. Class guideline DNVGLCG0152. Edition April 2016 Page 14 Plus extended fatigue analysis of ship details Section 4 SECTION 4 STRESS ANALYSIS OF STIFFENERFRAME CONNECTIONS

15 1 General This section applies to the longitudinal stiffenerframe connection details. A finite element analysis with a seminominal stress model (50 50 mm mesh) is required for stress calculations of the stiffenerframe connections. For all the other details reference is made to DNVGL CG 0127 for guidance on the finite element modelling of fine mesh model and partial ship model. 2 Seminominal stress model 2.1 Modelling For calculation of the stress level at the longitudinal stiffenerframe connections a seminominal stress finite element model should be made. The purpose of the seminominal model is to capture the local geometric stress flow and effect of cutouts, web frametoes and tripping brackets. The model also captures more accurately the shear stress distribution from longitudinal stiffeners to web frame. The stress results are used together with stress concentration factors to obtain the hotspot stress range for use in fatigue calculation. The model should have approximately 50 mm elements at the critical hotspots for each longitudinal stiffenerframe connection in side, inner side, bottom, inner bottom and hopper. The longitudinal extent of the model should be at least one frame spacing on each side of the target frame. Examples of models are shown in Figure 1 and Figure 2. The element types to be used are: Quad elements: 8 node Triangular elements: 6 node Beam elements: 3 node. The seminominal stress model may be included in the cargo hold model or submodelling may be used. The slot geometry should be modelled as described in [3], see Figure 3. It is important that the area around all hotspot locations is modelled with 50 mm size elements. In cases where it is impossible to create square 50 mm elements an aspect ratio of 4 should not be exceeded. If elements different from mm size are needed these elements should preferably be placed away from the hotspots i.e. at midspan of cutout opening. Away from the hotspots the mesh can gradually become coarser. The rounded corners of the slot should be modelled as square elements. Eccentric lugs are not to be modelled as eccentric but as inplane shell elements, and the plate thickness should not be increased to account for the overlapping plates. The effect of eccentric lug induced bending stresses will be captured by the stress concentration factors. The cutout should be modelled as at is on drawings with correct width and height. As a consequence will the distance from the longitudinal top flange to cutout edge increase with half the thickness of top flange. The stiffener is idealized with plate elements in the centre of the actual stiffener. Guidance on meshing of the slot geometry is given for all connection types in App.A [3]. It is important that the finite element mesh is similar in order to ensure correct hotspot stress. The loads are to be the same pressure loads as applied to the cargo tank model. If a submodelling analysis is to be used, the applied loads will also include either prescribed displacements or prescribed forces/ stresses. Class guideline DNVGLCG0152. Edition April 2016 Page 15 Plus extended fatigue analysis of ship details Section 5 SECTION 5 FINITE ELEMENT MODELLING OF STIFFENERFRAME CONNECTIONS

16 All types of stiffenerframe connections in the cargo area should be analysed using seminominal models. If the midhold target webframe does not include all types of connections then the remaining should be included in another model based on the midhold webframe model. This should also be done for all connections in the forward and aft cargo holds. Since a vessel does not have parallel body in the forward and aft tank area some simplification will be necessary. The remaining frames of the cargo tanks should be assessed based on a screening procedure described in [6]. Figure 1 FEmodel with 50 mm mesh, whole model Class guideline DNVGLCG0152. Edition April 2016 Page 16 Plus extended fatigue analysis of ship details Section 5 2.2

17 Section 5 Figure 2 FEmodel with 50 mm mesh, hopper tank Figure 3 Slot geometry with stress read out points Class guideline DNVGLCG0152. Edition April 2016 Page 17 Plus extended fatigue analysis of ship details

18 3.1 If the design of a longitudinal stiffenerframe connections is not found among the typical designs listed in App.A [4] a very fine mesh model (stress concentration model) should be made in order to establish the stress concentration factor of that particular design. The stress concentration models should be modelled according to the procedure given in this section. App.A [2] describes how the stress concentration factors are calculated while Sec.6 describes necessary documentation. 3.2 Two models are needed if additional stress concentration factors are to be made: mesh model: This model shall be used to predict the seminominal stress. t t mesh model: This model shall be used to calculate the hotspot stress. The stress concentration factor will be the stress ratio between the models. The models are made to simulate the behaviour of double side and double bottom. The models should have a vertical extent of 3 stiffeners, i.e. 4 stiffener spacing, and the longitudinal extent should be ½ frame spacing in both forward and aft direction. Two typical models are shown in Figure 4 and Figure 5. No cutouts for manholes should be included in the stress concentration models. Both models should include the typical hotspots for the target detail. Sec.2 Figure 1 shows typical hotspots to be assessed. The element types to be used are: Quad elements: 8 node Triangular elements: 6 node Beam elements: 3 node. Class guideline DNVGLCG0152. Edition April 2016 Page 18 Plus extended fatigue analysis of ship details Section 5 3 Stress concentration models

19 Section 5 Figure 4 FEModel with 50 mm mesh, SCF analysis Figure 5 FEModel with t t mesh, SCF analysis Class guideline DNVGLCG0152. Edition April 2016 Page 19 Plus extended fatigue analysis of ship details

20 For the t t mesh model the mesh density in the area of the web frame slots, including web stiffeners, is to be approximately the plate thickness. The mesh with plate thickness size should extend at least four elements in all directions at all relevant hotspots. Outside this area the mesh density may be increased to reduce the size of the model. The element aspect ratio should however not exceed 4. The eccentricity of the stiffener lug is to be included in the model and modelled according to Figure 8. Figure 6 shows an example of a stiffener lug detail modelled with t t mesh. Figure 6 Stiffener and lug details with web stiffener on top, t t mesh Figure 7 Stiffener and lug details with web stiffener on top, mesh Class guideline DNVGLCG0152. Edition April 2016 Page 20 Plus extended fatigue analysis of ship details Section 5 For the mm mesh model, the mesh density in the area of the web frame slots, including web stiffeners, is to be approximately mm but adjusted to fit the slot geometry. Outside this area the mesh size may be increased to reduce the size of the model. The element aspect ratio should however not exceed 4. Smooth corners are to be modelled as sharp corners and the eccentricity of the stiffener lug is to be ignored. An example of a detail is shown in Figure 7.

21 Section 5 Figure 8 Modelling of eccentric collar plate 4 Loads Three load effects are to be modelled in both models. LC1: External pressure LC2: Shear stress LC3: Axial load Each load effect will result in a stress concentration factor. In order to combine the stress concentration factors into one stress concentration factor, weighting of the different load effects should be used. The weighting is to be conducted by scaling the applied load according to the following criteria: 2 LC1: External pressure, nominal shear stress in the middle of the stiffener web of 100 N/mm. LC2: Shear stress, nominal stress in the middle of the web frame of 100 N/mm. LC3: Axial stress, nominal stress in the middle of the web frame of 100 N/mm. 2 2 The applied pressure and prescribed displacements should be used to obtain the nominal stress criteria in the areas indicated in Figure 9 (areas shaded in grey). The different load effects and boundary conditions are shown in Figure 10 through Figure 12. The forward and aft part of the finite element model should have symmetry condition describing the behaviour in a double side or double bottom. Class guideline DNVGLCG0152. Edition April 2016 Page 21 Plus extended fatigue analysis of ship details

22 Section 5 Figure 9 Nominal stress criteria Figure 10 Load application and boundary conditions for LC1, external pressure Class guideline DNVGLCG0152. Edition April 2016 Page 22 Plus extended fatigue analysis of ship details

23 Section 5 Figure 11 Load application and boundary conditions for LC2, shear stress Figure 12 Load application and boundary conditions for LC3, axial stress 5 Stress read out from FE models 5.1 The maximum principal element stress within ±45º of the normal to the weld should be used for the analysis. As a conservative approach the absolute maximum principal stress could be used regardless of direction to the weld. 5.2 The seminominal stress should be the maximum principal membrane stress at the considered hotspot location. The stress read point on the seminominal model is to be at the hotspot node location. Among all the elements that have a result at the considered hotspot node it is the element result with maximum principal stress value that should be used for damage calculation. No averaging of nodal stresses is allowed. See Figure 3, Figure 13 and Figure 14 for illustration. If necessary the membrane stress should be calculated using the surface stresses at upper and lower surfaces. Class guideline DNVGLCG0152. Edition April 2016 Page 23 Plus extended fatigue analysis of ship details

24 Section 5 Figure 13 Example of stress read out point from mm mesh model Figure 14 Example of maximum principal membrane stress value among adjacent elements 5.3 The maximum seminominal stress value for a certain hotspot node may be found at different elements among the various load conditions. The seminominal stress values should not necessarily be taken from the same element in all load conditions. For a certain hotspot the result values should be found at the same node position but they might result from various elements connected to that node. 5.4 The stress read out points in the stress concentration model is to be at a distance t/2 from the hotspot. The principal element stresses at t/2 should be multiplied with a factor 1.12 to calculate the hotspot stress. This Class guideline DNVGLCG0152. Edition April 2016 Page 24 Plus extended fatigue analysis of ship details

25 Figure 15 Example of stress read out point from t t mesh model 5.5 When extracting stress results from the stress concentration model there could for a given hotspot be many possible t/2nodes. The maximum principal stress value at position t/2 away from the hotspot location should be used as the relevant hotspot stress. All t/2positions will have to be assessed with regard to stress level in combination with principal stress direction in order to locate the maximum relevant stress value. See Figure 16 for illustration. 5.6 The maximum principal stress value at t/2position may occur at different positions for the three considered load cases. Despite possibly different location of maximum stress location the stress read out node in the stress concentration model should for a given hotspot remain the same for all three load cases: axial, shear and pressure load. The node position giving the absolute maximum value among all three load cases should be used when extracting stress for all three load cases. Note that both upper and lower element surface should be checked in order to find the maximum principal stress. 5.7 At the stress read out node the finite element model will report a result for both upper and lower surface and also principal stress in two directions. The principal stress direction that is within 45 degrees to the weld normal should be used as relevant hotspot stress. If both principal stresses are both at 45 degrees to the weld normal one should choose the maximum of the two directions. Both upper and lower surface stress should be reported for all three load cases. The surface resulting in the largest total stress when summing up the contributions from the load cases should be used as basis for calculation of stress concentration factor. 5.8 For a specific hotspot location the stress read out procedure for the stress concentration model can be summarized as follows: Locate all possible t/2nodes with principal stress direction within 45 degrees to weld normal (t/2 nodes: nodes with a distance of t/2 from the hotspot). Find the t/2node with the largest principal surface stress among all load cases. Class guideline DNVGLCG0152. Edition April 2016 Page 25 Plus extended fatigue analysis of ship details Section 5 is one of the two stress extrapolation procedures given in DNVGL CG See Figure 15 and Figure 16 for illustration.

26 Figure 16 Possible stress read out nodes at t/2 position 5.9 In the stress concentration model the collar plate is modelled with eccentricity. In the area where the collar plate overlaps with the web plate, the stress is normally reduced due to the increase effective plate thickness. Some hotspots are located on the boundary between web plate and collar plate. For these hotspot the stress results in the overlapping area need not be considered. These stress values are assumed not to give rise to fatigue crack growth. Figure 16 illustrates the overlapping area Example of relevant elements to consider for identification of possibly maximum principal stress values at t/2 position are shown in Figure 17 for the hotspots of detail type T201. For the hotspots #101, #107, #202, #206, #207 and #208 is the marked element the only relevant to consider regarding t/2stress values. For hotspot #104, #105 and #106 which are located at the slit opening edge, several element along the edge needs to be considered marked with arrows on Figure 17. Hotspot #102, #103, #204 and #205 are located adjacent to the area where the lug plate and the webplate are overlapping each other. Elements that are inside this overlapping area will not be necessary to consider when locating the relevant t/2stress result. Normally these elements will have smaller stress values and stresses in this area will not contribute to crack growth in real structures. Class guideline DNVGLCG0152. Edition April 2016 Page 26 Plus extended fatigue analysis of ship details Section 5 Perform stress read out of both first and second principal stress on upper and lower surface at the same node for all three load cases. For a certain load case locate the largest principal stress among the two principal directions (P1 or P2) for both upper and lower surface. To decide on which surface to use calculate the sum of the three load cases for both upper and lower surface. The surface with the largest total stress should be used for calculation of the SCF. The SCF is calculated according to the expression in App.A [2.2].

27 Section 5 Figure 17 Elements to consider in stress read out 5.11 The hotspots on the edge of the slit opening i.e. hotspot #104, #105 and #106 are located in base material. Stress read out on hotspots in base material should be performed at the hotspot. No extrapolation using t/2 stress values should be performed and stress results on the opening edge should be used. The use of dummy beam elements at slit opening edge is not allowed. It is found that significant bending may occur at the slit opening hotspots. Since the dummy beam element does not capture bending stress across the plate thickness they are not allowed to use. Note that the maximum stress position along the slit opening most likely will vary for the various load cases. The position with the largest value among the load cases should be used for stress read out for all three load cases. Class guideline DNVGLCG0152. Edition April 2016 Page 27 Plus extended fatigue analysis of ship details

28 Section 5 Figure 18 Possible stress read out nodes at slit opening edge 6 Screening procedure The below screening procedure should be used to assess the fatigue capacity of the webframes that are not modelled using a mm element mesh. The screening should be based on the nominal stress level of the cargo hold model. The screening analysis should follow the following steps: 1) Establish a scaling factor for each stiffenerframe connection in each load case. The scaling factors should be based on maximum element average principal stress from the cargo hold model at the reference connection and at the target connection. The webframe that has been modelled with mm element mesh is defined as the reference webframe. The scaling factor, fs, is defined as the ratio between the average value of the absolute principal stress level of four neighbouring elements at the relevant stiffenerframe connection in the target frame and the reference frame, Figure 18: 2) Establish hotspot stress at target stiffenerframe connection by multiplying the scaling factor with the hotspot stress of the relevant hotspot location at the reference stiffenerframe connection. The reference hotspot stress is established by use of the seminominal model and stress concentration factors according to Sec.3 and App.A [4]. Fatigue damage is calculated for the relevant hotspot location at the target stiffenerframe connection. The above procedure is repeated for all hotspots at all stiffenerframe connections in all webframes in the forward, aft and amidship cargohold models. 3) 4) Class guideline DNVGLCG0152. Edition April 2016 Page 28 Plus extended fatigue analysis of ship details

29 Note that when different connections having similar mm mesh are compared it is important to account for the correct SCF giving the relevant reference hotspot stress. If the connections in the target and reference webframe would have different mm mesh the screening procedure should be used with care. In such case, the mm element mesh for the reference connection may be modified to reflect the geometry of the target connection. Figure 19 Neighbouring elements for screening average stress Class guideline DNVGLCG0152. Edition April 2016 Page 29 Plus extended fatigue analysis of ship details Section 5 The above screening procedure is only applicable when comparing similar connections in the target and the reference webframe. Since the mm mesh is similar for many different connections it will be possible to use the screening procedure for a number of different connections even if the two connections in the target and the reference webframe are somewhat different.

30 1 FEmodels All finite element models should be documented with plots clearly showing the mesh at the various details. The model extension and a thorough description of the applied loads and boundary conditions should be documented. Element types should be reported. A description of the analysis flowchart should be included as documentation for to explain the analysis flow. 2 Stresses The seminominal stress in the web frame should be documented with stress plots showing clearly the stress distribution at all stiffeners. For validation of the predicted fatigue damages all stress values for all locations should be reported in numerical format. 3 Fatigue calculations The fatigue calculations should be reported including the following items: Choice of SNCurves Fraction of time in each loading condition Fraction of time in noncorrosive/corrosive environment Number of stress cycles Mean stress correction factors Total fatigue damage or fatigue life prediction Stress concentration factors. 4 Stress concentration factor analysis Analysis of additional stress concentration factors should be thoroughly documented. A detailed description of the longitudinal stiffenerframe connection geometry is to be provided together with a description and plots of both the t tmodel and the mesh finite element models. The applied loads and boundary condition should also be documented, and plots clearly showing the element mesh around the stiffenerframe connection should be provided. The nominal stress level and the hotspot stress should be documented with plots for all loadcases in both models. The stresses used in calculation of the stress concentration factor are to be reported together with the resulting stress concentration factor. Class guideline DNVGLCG0152. Edition April 2016 Page 30 Plus extended fatigue analysis of ship details Section 6 SECTION 6 DOCUMENTATION OF PLUS ANALYSIS

31 1 General 1) 2) Det Norske Veritas, Finite element analysis of large scale fatigue test of stiffener web frame connections, DNV report No , T. Lindemark, november 2007 Lotsberg, Inge (et.al) Fatigue Capacity of Stiffener to Web Frame Connections, OMAE , Honolulu, Hawaii, USA, 31 may5 june Class guideline DNVGLCG0152. Edition April 2016 Page 31 Plus extended fatigue analysis of ship details Section 7 SECTION 7 REFERENCES

32 1 General This section describes the procedure of how to establish stress concentration factors for longitudinal stiffenerframe connection details. [4] lists stress concentration factors for the critical hotspots of typical stiffenerframe connections. If a particular detail is not listed among the typical ones in [4] then the user should follow the procedure in [2] for calculating the stress concentration factors. 2 Establish stress concentration factors 2.1 In order to calculate the stress concentration factor a seminominal model with 50 50mm mesh and a stress concentration model with t tmesh should be made. The finite element modelling and analysis should be performed according to the procedure described in Sec When extracting stress results the maximum element principal stresses should be used and not the node averaged stresses. The stress read out points should be located: At t/2 from the intersection line in the t tmesh model, see Sec.5 Figure 16. At the node at hotspot in the 50 50mm mesh model, see Sec.5 Figure 13. Note that the stress at t/2 should be multiplied with an extrapolation factor of 1.12 to find the hotspot stress. See DNVGL CG 0129 for details on extrapolation procedures. The maximum principal stress within ±45º of the normal to the weld should be used for the analysis. This requirement will in most cases decide whether first or second principal stress should be used. For some hotspots the position of maximum element principal stress could vary among the three load cases. Note that the stress read out position for a given hotspot should be the same for all load cases. The position giving the maximum stress value among all load cases should be used for stress read out in all load cases. The SCFs are calculated as the ratio of the principal stresses between the seminominal model and the fine mesh model. The sum of absolute principal stresses at read out position is to be used for both models as described in the formula: σ1hs = hotspot stress in first principal direction σ2hs = hotspot stress in second principal direction σ1t/ σ2t/ Class guideline DNVGLCG0152. Edition April 2016 Page 32 Plus extended fatigue analysis of ship details Appendix A APPENDIX A STRESS CONCENTRATION FACTORS

33 = maximum first principal element stress at t/2 position Appendix A σ1t/2 σ2t/2 σm,1 σm,2 = maximum second principal element stress at t/2 position = seminominal principal membrane element stress at hotspot node = seminominal principal membrane element stress at hotspot node. The checked hotspots should normally include the hotspots marked in Sec.5 Figure The stress concentration factor for hotspot #102 and #103 see Sec.2 Figure 1, should be based on effective hotspot stress according to the expression below: The effective hotspot stress reduces the bending component of the surface stress with a factor of 0.6. Effective stress is calculated according the following expression: = membrane stress = bending stress. 3 Example of establishing stress concentration factors 3.1 This example illustrates how to establish stress concentration factors for a Tshaped longitudinal through a slot with lug connection. The example is based on finite element analysis as described in Sec.5 [4] and shows how to weight the different load cases and stresses, and how to establish the stress concentration factors. The stresses from the finite element analysis are found in Table 1. The magnitude of the sum of the stresses in the t tmesh model is used to predict the criticality of the hotspots. The critical hotspots will vary from detail to detail dependent on the different detailed design. The weighted stress concentration factors are shown in Table 1. The weighted stress concentration factors are established based on the columns marked with sum, which is basically the sum of the different load cases as illustrated in the formula above. The weighted stress concentration factors should be used in the fatigue calculations. 3.2 Flowchart The flowchart, Figure 1, shows the main steps of the procedure to be followed when establishing a stress concentration factor, reference is given to the indicated sections for detailed guidance. Class guideline DNVGLCG0152. Edition April 2016 Page 33 Plus extended fatigue analysis of ship details

34 Appendix A Figure 1 Flowchart 3.3 Loads To simulate the stress flow in a double side or bottom the two models needed for calculation of stress concentration factors should both be subjected to pressure, axial and shear loads as described in Sec.5 [5]. 3.4 Seminominal model A mm mesh model should be made according to the requirements given in Sec.5 [3]. The model is used to calculate seminominal stresses. Figure 2 shows the membrane principal stress distribution at the stiffenerframe connection and the stress read out position at the example hotspot #102 for the pressure load case. Table 1 lists stress values for all load cases. Figure 2 Seminominal membrane principal stress Class guideline DNVGLCG0152. Edition April 2016 Page 34 Plus extended fatigue analysis of ship details

35 A t tmm mesh model should be made according to the requirements given in Sec.5 [4]. The model is used to calculate the hotspot stresses. Figure 3 shows the hotspot principal stress distribution at the stiffenerframe connection and the stress read out position at the example hotspot #102 for pressure load case. Table 1 lists hotspot stresses for all load cases. Note that the stress read out point on the t tmesh model is at t/2 position away from the hotspot location, and that the maximum element principal stress within ±45 degree of the weld normal should be used. Figure 3 Hotspot surface principal stress 3.6 Stress concentration factor calculation The SCFs are calculated as the ratio between the membrane principal stresses from the seminominal and the surface principal stress from the fine mesh model. The sum of the absolute principal stresses from pressure, axial and shear loads are to be used for both models as described in the formula: By using the stress values for the example hotspot is the following SCF (Kg) calculated: Class guideline DNVGLCG0152. Edition April 2016 Page 35 Plus extended fatigue analysis of ship details Appendix A 3.5 Fine mesh model

36 According to [2.3] the example hotspot #102 in Figure 3 should be based on effective stress with a reduction factor of 0.6 on the bending stress component. This reduction is accounted for in the above example calculation. The effective stress of HS102 for LC1 will be: σeff,lc1,hs102 = σb σm = = 683 Table 1 summarize the SCF (Kg) values for all hotspots for the example detail shown in Figure 2. Table 1 Stresses for SCF calculation and SCF (Kg) values Seminominal stress model Hotspot t t model LC1 Pressure LC2 Shear LC3 Axial Sum LC1 Pressure LC2 Shear LC3 Axial Sum # = 1169 Nominal stress SCF (Kg) Stress concentration factors for typical longitudinal end connection details 4.1 The stress concentration factors listed below covers typical stiffenerframe connections found in ship structures. Figure 4 shows possible hotspot locations and the numbering system. Table 2 shows the dimensions of the details used in calculation of the tabulated SCFs. Stress concentration factors for details without web stiffener, with web stiffener, and with backing bracket and soft toe are listed in Table 4. Stress concentration factors for hotspots on the web stiffener are listed in Table 5. If the detail to be checked differs significantly from the typical details in Table 2 separate analysis should be conducted according to [2]. Class guideline DNVGLCG0152. Edition April 2016 Page 36 Plus extended fatigue analysis of ship details Appendix A Note that fine mesh stress values taken at t/2 position should be multiplied by an extrapolation factor of 1.12 in order to obtain the hotspot stress.

37 Appendix A Figure 4 Numbering of possible hotspots Class guideline DNVGLCG0152. Edition April 2016 Page 37 Plus extended fatigue analysis of ship details

38 Appendix A 4.2 Table 2 Dimensions of calculated stiffenerframe connections Class guideline DNVGLCG0152. Edition April 2016 Page 38 Plus extended fatigue analysis of ship details

39 Appendix A Class guideline DNVGLCG0152. Edition April 2016 Page 39 Plus extended fatigue analysis of ship details

40 Appendix A Class guideline DNVGLCG0152. Edition April 2016 Page 40 Plus extended fatigue analysis of ship details

41 Appendix A Table 3 Dimensions of top stiffener connections Web stiffener with high scallop Web stiffener with keyhole Web stiffener and brackets Class guideline DNVGLCG0152. Edition April 2016 Page 41 Plus extended fatigue analysis of ship details

42 Geometry Hotspot Kgfactor with web stiffener and high scallop Kgfactor with soft toe and bracket Kgfactor without web stiffener * * Class guideline DNVGLCG0152. Edition April 2016 Page 42 Plus extended fatigue analysis of ship details Appendix A Table 4 Stress concentration factors for stiffenerframe connections

43 Hotspot * * Kgfactor with soft toe and bracket Kgfactor without web stiffener Class guideline DNVGLCG0152. Edition April 2016 Page 43 Plus extended fatigue analysis of ship details Appendix A Geometry Kgfactor with web stiffener and high scallop

44 Hotspot Kgfactor with soft toe and bracket Kgfactor without web stiffener * * * * 2.23* 2.21* * * * * 2.24* 2.21* Class guideline DNVGLCG0152. Edition April 2016 Page 44 Plus extended fatigue analysis of ship details Appendix A Geometry Kgfactor with web stiffener and high scallop

45 Hotspot Kgfactor with soft toe and bracket * * * 2.94* Kgfactor without web stiffener * * * 0.81* Class guideline DNVGLCG0152. Edition April 2016 Page 45 Plus extended fatigue analysis of ship details Appendix A Geometry Kgfactor with web stiffener and high scallop

46 Hotspot Kgfactor with soft toe and bracket Kgfactor without web stiffener Class guideline DNVGLCG0152. Edition April 2016 Page 46 Plus extended fatigue analysis of ship details Appendix A Geometry Kgfactor with web stiffener and high scallop

47 Hotspot Kgfactor with soft toe and bracket Kgfactor without web stiffener * based on effective stress Class guideline DNVGLCG0152. Edition April 2016 Page 47 Plus extended fatigue analysis of ship details Appendix A Geometry Kgfactor with web stiffener and high scallop

48 Geometry Hotspot Kgfactor with web stiffener for high scallop Kgfactor with web stiffener for keyhole Kgfactor with web stiffener and brackets T T T T T T T T Seminominal finite element mesh of stiffenerframe connections Figure 5 and Figure 6 show the seminominal element mesh that is used to calculate the stress concentration factors in Table 4 and Table 5. It is important that a similar element mesh configuration is used for the respective stiffenerframe connections in the seminominal webframe model. Class guideline DNVGLCG0152. Edition April 2016 Page 48 Plus extended fatigue analysis of ship details Appendix A Table 5 Stress concentration factors for web stiffener connections to longitudinal

49 Appendix A Class guideline DNVGLCG0152. Edition April 2016 Page 49 Plus extended fatigue analysis of ship details

50 Appendix A Figure 5 Seminominal element mesh at web cutout Class guideline DNVGLCG0152. Edition April 2016 Page 50 Plus extended fatigue analysis of ship details

51 Appendix A Figure 6 Seminominal element mesh at web stiffener Class guideline DNVGLCG0152. Edition April 2016 Page 51 Plus extended fatigue analysis of ship details