Design and Analysis of a GFRP Sandwich Panel Subjected to Anchor Point Loading Conditions

8/11/2019 Design and Analysis of a GFRP Sandwich Panel Subjected to Anchor Point Loading Conditions

1/53

Design and Analysis of a GFRP Sandwich Panel subjected to

Anchor Point Loading Conditions

A Project Report

BY

SRIVATSAN A.V 2008H141056P DESIGN ENGINEERING

Prepared in partial fulfillment of the

PRACTICE SCHOOL-2 COURSE

AT

VESTAS TECHNOLOGY R&D, CHENNAI

A PRACTICE SCHOOLII STATION OF

BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, PILANI

(JUNE, 2010)


2/53

II

BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE

PILANI (RAJASTHAN)

Practice School Division

Station: Vestas Technology R&D

Centre: Chennai

Duration: 06th

January 2010 to 18th

June 2010

Date of Start: 06th January

Date of Submission: 10th

June 2010

ID No. /Name: 2008H141056P/ Srivatsan.A.V

Discipline: Design Engineering

Name and Designation Of the expert: Mr. Dhanabal Raguraman,

Deputy Manager, Nacelle & Hub

Name of the PS Faculty: Mr. K. Venkataraman

Keywords: Composite materials, Sandwich structures, Nacelle covers,

Anchor points, GFRP

Project Areas: Composite Materials, Design & Analysis

Signature of Student Signature of PS Faculty

Date Date


3/53

III

BIRLA INSTITUE OF TECHNOLOGY & SCIENCE

PILANI (RAJASTHAN)

PRACTICE SCHOOL DIVISION

Response Option Sheet

Station: Vestas Technology R&D

Centre: CHENNAI

ID No. & Name: 2008H141056P,SRIVATSAN.A.V

Title Of the Project: DESIGN OF GFRP COMPOSITE SANDWICH PANEL

SUBJECTED TO ANCHOR POINT LOADING CONDITIONS

Usefulness of the project to the on-campus courses of study in various disciplines. Projectshould be scrutinized keeping in view the following response options. Write Course No.And Course Name against the option under which the project comes.

Refer Bulletin for Course No, and Course Name.

Signature of Student Signature of Faculty

Code No. Response Options Course No. & Name

1. A new course can be designed out of thisproject

NO

2. The project can help modification of thecourse content

Of some of the existing courses.

YES

MATERIALSTECHNOLOGY&TESTING

3. The project can be used directly in some ofthe existing Compulsory Discipline

courses (CDC) /Disciplines CompulsoryDiscipline courses (DCOC) / Emerging

Area (EA) etc. Courses

YES

FRACTURE MECHANICS

4. The project can be used in preparatorycourses like Analysis and Application

Oriented Courses (AAOC)/ EngineeringScience (ES)/ Technical Art (TA) and Core

Courses.

NO

5. This project cannot come under any of the

above-mentioned options as it relates to theprofessional work of the host organization.

NO


4/53

IV

ACKNOWLEDGEMENTS

I offer my sincere thanks to the management of Vestas Technology R&D, Chennai for giving

me this opportunity to work on this project as part of the PS-2 program.

I would like to thank Mr. Etekamba Okon Willie, Mr. Siva Kumar, Mr. Gopan P.R and Mr.

Anu Sharma for giving me the opportunity to work on this project in the Nacelle & Hub team.

I would like to thank Mr. Dhanabal Raguraman for assigning me this project and mentoring

me during the course of this project by giving vital inputs throughout the course of my

project.

I thank Mr. Padmanathan Sudalaimuthu for giving me vital inputs and support which helped

me in the course of this project.

I thank Mr. Achuthan Babu, Mr. Manikandan Rajagopal and Mr. N Guruguhan of Blades

division for giving me valuable inputs and support in the course of this project.

I thank Mr. Umashankar Keecheril Ayyappan and Mr. Ravi Kishore Chaganti of Tower and

Structures division for their support in the course of this project.

I sincerely thank Mr. K. Venkataraman (PS faculty) and BITS PS division for giving me

valuable inputs for the completion of the project and being a source of constant support and

encouragement.

I would also like to thank other members in Nacelle & Hub team namely, Mr. Raghavendra

Babu Karanam, Mr. Gnanasagar Vaithilingam, Mr. Venkatesh Krishnan, Mr. Pradeep B, Mr.

Amalan Paul Samuel, Mr. Mohanraj Kathiresan, Mr. Amar Kumar Sinha, Mr. Pravin Savant

for guiding me during the project by giving valuable suggestions.

I also thank my PS mate Mr. Jagan Mohan Ponnada, for supporting me during the course of

the project.


5/53

V

ABSTRACT

The nacelle of a wind turbine contains the prime components of a wind turbine generator.

Due to weight considerations, the nacelle covers are made up of composite sandwich

structures. During service the personnel access the roof of these nacelles and they hook

their safety harness to a special component on the cover called Anchor Points to prevent

an unexpected fall. In the case of an accidental fall, the reliability of these anchor point

means life or death of a person. This present work aims to design a suitable structural

configuration of a sandwich panel, exclusively for the suddenly applied loading

conditions that is offered by the anchor points. International standards and certifying

bodies demand that the panel to qualify a load twenty four times the load offered by the

actual weight of a person.

In this work the theoretical design and Finite Element analysis of the sandwich panel is

done. The design is based on specific loading conditions given by Anchor Points that the

panel must meet prior to failure under the load. Selection of materials to be used for the

facing and core are done based on given requirements. With the materials chosen, the

facing sheet and the core are analyzed and response is studied and checked for failure by

using the multi point constraint option in ANSYS software. This work will aid in the

design of the nacelle cover.


6/53

VI

LIST OF ABBREVIATIONS

CIM Continuous Improvement Management

AP Anchor Points

WTG Wind Turbine GeneratorPFA Personnel Fall Arrest

SA Safety Alert

ECR Engineering Change Request

ECO Engineering Change Order


7/53

VII

List of Figures

Figure 1.1 Wind turbine layout 2

Figure 2.1 Anchor Points on a nacelle roof 4

Figure 2.2 Anchor points arresting a fall 5

Figure 4.1 Schematic illustrations of common composite reinforcements 8

Figure 4.2 Loads on a plate 13

Figure 4.3 z-coordinates of the individual layers through a laminate 13

Figure 4.4 Sandwich Panel loading 15

Figure 5.1 Calculation of Ply properties 19

Figure 5.2 Calculation of Laminate Properties 19

Figure 6.1 Vacuum Assisted Resin transfer Molding 20

Figure 7.1 Direction of Anchor Point Loading 21

Figure 8.1 Simply supported Boundary Condition 23

Figure 9.1 Meshing and Boundary conditions 28

Figure 9.2 Deflection observed 29

Figure 9.3 Facing Stress 29

Figure 9.4 Core Shear Stress Without edge effects of Layer 2 30

Figure 9.5 Anchor Point Loading 31

Figure 9.6 MPC-Meshing and Loading conditions 32

Figure 9.7 Equivalent Stress in layer 1 33




Figure 9.11 Stresses along the X direction 35

Figure 9.12 Shear Stress in layer 9 Core 35


8/53

VIII

List of Tables

Table 5.1 Properties of different core 18

Table 5.2 Properties of the Fiber 18

Table 8.2 Output Parameters 25

Table 8.3 Partial safety factors 25

Table 8.4 Check for design loads versus material's resistance 25

Table 8.5 Validation theoretical model 26


9/53

IX

Nomenclature

a = Panel length

A = Area of applied load

b = Beam width

D = Panel bending stiffness

EC= Compression modulus of core

Ef= Modulus of elasticity of facing skin

F = Maximum shear force

GC= Core shear modulus - in direction of applied load

h = Distance between facing skin centres

kb= Beam - bending deflection coefficient

kS= Beam - shear deflection coefficient

l = Beam span

M = Maximum bending moment

P = Applied load

Pb= Critical buckling load

q = Uniformly distributed load

S = Panel shear stiffness

tC= Thickness of coretf= Thickness of facing skin

V = Panel parameter (used for simply supported plate)

d = Calculated deflection

sC= Core compressive stress

sCR= Critical facing skin stress

sf= Calculated facing skin stress

tC= Shear stress in core

m = Poissons Ratio of face material

l = Bending correction factor for Poissons Ratio effect


10/53

X

CONTENTS

ACKNOWLEDGEMENTS IV

ABSTRACT V

LIST OF ABBREVIATIONS VILIST OF FIGURES VII

LIST OF TABLES VIII

NOMENCLATURE IX

1. COMPANY PROFILE 1

1.1.DEPARTMENT PROFILE 1

1.2.WIND TECHNOLOGY 1

1.3.

MAIN COMPONENTS OF A WIND TURBINE 2

1.3.1.

ROTOR 2

1.3.2. TOWER 2

1.3.3. NACELLE 3

2. INTRODUCTION 4

2.1.PURPOSE 5

3. LITERATURE SURVEY 6

4. COMPOSITE MATERIALS 7

4.1.BASIC COMPOSITE THEORY 7

4.2.TYPES OF COMPOSITES 8

4.3.POLYMER MATRIX COMPOSITES 9

4.4.CLASSICAL LAMINATION THEORY OF COMPOSITES 11

4.5.SANDWICH STRUCTURES 14

4.5.1. STRUCTURAL DESIGN CRITERIA 15

4.5.2. FACING/SKIN DESIGN 16

4.5.3.

CORE DESIGN 16

5. MATERIAL SELECTION 17

5.1.CORE 17

5.2.LAMINATES 18

5.2.1. FIBRES 18


11/53

XI

5.2.2. RESIN SYSTEM 18

5.2.3. CALCULATION OF PLY AND LAMINATE PROPERTIES 19

6. MANUFACTURING PROCESS 20

7. DESIGN CONSIDERATIONS 21

7.1.PARTIAL SAFETY FACTOR METHOD 22

8. THEORETICAL CALCULATION 23

8.1.VALIDATION OF THE THEORETICAL MODEL 26

9. FEA VALIDATION 27

9.1.PREPROCESSINGMODEL VALIDATION 27

9.2.POST PROCESSING - MODEL VALIDATION 28

9.3.

PREPROCESSINGMULTI POINT CONSTRAINT 31

9.4.

POST PROCESSINGMULTI POINT CONSTRAINT 33

9.5.RESULTS AND DISCUSSION 36

10.CONCLUSION 37

11.REFERENCE 38

12. APPENDIX 40D

LOADINGS 12

11.1.1.TE JOINT ANALYSIS 12

11.2. S P A R C A P T O S H E L L J O I N T

1 3

11.2.1.GEOMETRY AND LOADINGS 13

11.2.2.SPAR CAP TO SHELL JOINT ANALYSIS 23

12.COHESIVE ZONE MODELING 15

13.FRACTURE MECHANICS BASED METHODS 17

17

17


12/53

1

1. Company Profile

Vestas, a leading wind turbine manufacturer, has its research and development centre at

Chennai. The Chennai centre is on the IT corridor, spread over 60,000 sq ft, has started

with more than 140 employees, which number will go up to 500 engineers in the next

four years. Vestas Indian subsidiary is headquartered in Chennai a long the IT corridor,

close to where the R&D centre is located. Vestas Chennai R&D centre will support the

global engineering research and development activities, working all over the value chain.

The engineers work on the most attractive projects in the value chain mechanical,

aerodynamics, material research and electronics.

1.1 Department Profile

NHD is responsible for the layout of the Nacelle and Hub with focus on Volume of

Control, Volume of Activity and Weight balance. NHD is responsible for the Internal

Crane and Covers plus all auxiliary components. It is responsible for the conditioning

(Heating/Cooling, Ventilation and Air Condition) and corrosion protection of the Nacelle

and its systems and components. NHD is responsible for integrating systems, modules

and components in future WTG projects for the Nacelle and Hub.

1.2 Wind Technology

Wind is a form of solar energy. Winds are caused by the uneven heating of the

atmosphere by the sun, the irregularities of the earth's surface, and rotation of the earth.

Wind flow patterns are modified by the earth's terrain, bodies of water, and vegetation.

Humans use this wind flow, or motion energy, for many purposes: sailing, flying a kite,

and even generating electricity.

The terms wind energy or wind power describes the process by which the wind is used to

generate mechanical power or electricity. Wind turbines convert the kinetic energy in the

wind into mechanical power. This mechanical power can be used for specific tasks (such

as grinding grain or pumping water) or a generator can convert this mechanical power

into electricity.


13/53

2

1.3 Main Components of a Wind turbine

1.3.1 Rotor

Structurally, the rotor of a wind turbine consists of a number of subsystems. Based on the

definition that the rotor comprises all rotating parts of the unit outside the nacelle, thesesubsystems are the rotor blades, the hub, and the blade pitch mechanism, all three of

which are largely autonomous components with regard to their design, their operation

and the manufacturing techniques used. The rotor hub and the blade pitch mechanism

represent traditional mechanical engineering. Depending on the design, the blade pitch

system and its control system are only partly rotor components. Rotor blade technology is

associated more with lightweight aeronautical engineering than with conventional

mechanical engineering.

Figure 1.1 Wind turbine layout

1.3.2 Tower

The high tower is an essential component of the horizontal-axis turbine, a fact which can

be both an advantage and a disadvantage. The costs, which can amount to up to 20%of

the overall turbine costs, are, of course, disadvantageous. As the height of the tower

increases, transportation, assembly and erection of the tower and servicing of the

components also become increasingly more difficult and costly. On the other hand, the

specific energy yield of the rotor also increases with tower height.


14/53

3

1.3.3 Nacelle

In almost all turbines, the components of the mechanical drive train and of the electric

generator are housed in a closed nacelle. Some smaller turbines make do without it. A

completely closed housing could become redundant in case of a complete integration ofthe drive train components, for example by mounting the rotor bearings directly on the

gearbox. After all, the nacelle does represent a considerable cost factor. On the other

hand, many practical reasons speak for a closed nacelle, particularly in large turbines.

The nacelle cover is made of fiberglass. Hatches are positioned in the floor for lowering

or hoisting equipment to the nacelle and evacuation of personnel. The roof section is

equipped with wind sensors and skylights which can be opened from inside the nacelle to

access the roof and from outside to access the nacelle. The nacelle cover is mounted on

the girder structure.

The project mainly focuses on this nacelle cover and the anchor points on them, which

arrests the fall of a person.


15/53

4

2. Introduction

The main function of the Anchor points present on the top of the nacelle covers is to

arrest the fall of the service personnel. Service personnel while accessing the roof for

carrying out the service on the wind turbines, hook on their personal protective

equipment on to metallic members attached to the nacelle roofs. Generally the Anchor

devices are made of sheet metal and a hole is drilled in to the composite cover and

mounted. The attachment between the anchor points and the composite cover can be

called as bolted joints. There are many ways in which the failures may happen in the case

of bolted joints such as, the bearing failure, fastener pull out, shear out failure and so on.

The load bearing capacity of the composite structure used should be sufficient to prevent

these failures. The nacelle covers are made of sandwich structures, with a foam coresandwiched between two facings, made of composite laminates with specific fiber

orientation and stacking sequence. The laminates are of equal thickness to ensure uniform

flexure behavior of the sandwich structure.

The entire top portion of the nacelle roof contains many such anchor points. This is so

designed to ensure that service personnel have access to the complete part of the nacelle

during service. The following picture effectively explains the number of the anchor

points in a wind turbine nacelle.

Figure 2.1 Anchor Points on a nacelle roof


16/53


17/53

6

3. Literature Survey

Sandwich composites are widely employed in modern mechanical design, not only in the

field of aeronautical constructions, where they have initially been developed, but also in

the fields of land transports and wind energy. Almost the entire visible components of a

wind turbine excepting the tower are made from composite materials. The facing sheets

carry axial tensile and compressive stresses and the core material sustains shear and

compressive stresses normal to the panel. The core to facing bonding adhesive rigidly

joins the facing sheets and core materials and allows them to act as a single entity with a

high torsional and bending rigidity [25]. Because of their main features, such as the high

flexural resistance and stiffness [17], the high impact strength [18,19], the high corrosion

resistance [20] and the low thermal and acoustics conductivity [2122], sandwich

structures are in fact preferred over conventional materials in various industrial

applications. Although large number of research projects have been performed by various

authors, the design of structural elements made from sandwich composites is often a

difficult task. This is mainly because a reliable strength prediction needs the preliminary

knowledge of the mechanical behaviour of skins and core, as well as of the peculiar

damage mechanisms [2324] and failure criteria that can be used under a complex

loading. The goal of the present work is to come up with the design of a sandwich panel

which can take a load of 36kN. The structural configuration of the sandwich panel is to

be used in the design of nacelle covers


18/53

7

4. Composite Materials

4.1 Basic Composite Theory

A Composite Material is a materials system composed of a mixture or combination of

two or more micro- or macro-constituents that differ in form and chemical composition

and which are essentially insoluble in each other. Composites involve two or more

component materials that are generally combined in an attempt to improve material

properties such as stiffness, strength, toughness, etc.; the resulting properties are largely

dependent on the distribution, relative amounts and geometries of the constituents.

Composites consist of two or more phases that are usually processed separately and then

bonded, resulting in properties that are different from those of either of the component

materials. Polymer matrix composites generally combine high-strength, high-stiffness

fibers (graphite, kevlar, etc.) with low-density matrix materials (epoxy, polyvinyl, etc.) to

produce strong & stiff materials that are lightweight. Laminates are generally built up

from multiple layers of lamina; the fibers within each lamina are generally parallel, but

laminates usually contain lamina with their fibers oriented in various directions. Each

lamina is an anisotropic layer with properties varying as a function of fiber angle.

Loading along the fibers (longitudinal) is modeled as Isostrain while loadingperpendicular to the fibers (transverse) is modeled as Isostress; these two directions

generally represent the extremes in material behavior. Fiber and matrix material property

data can be used to predict/approximate the properties of laminated composites using the

Rule of Mixtures. In this investigation, the elastic modulus of composites loaded at

various angles with respect to the fiber direction will be predicted, tested and discussed.

The burn-off method will be applied to determine the fiber volume of the composites

being investigated, and ultimate strength and elastic modulus results will be compared

with those of metals and polymers.


19/53

8

4.2 Types of composites

The 3 basic types of composites are identified as

1) Particle-Reinforced (Aggregates)

2) Fiber-Reinforced (Continuous Fiber or Chopped Fiber)

3) Natural Composites (Examples: Wood and Bone)

Figure 4.1 Schematic illustrations of common composite reinforcements

Today, the most common man-made composites can be divided intothree main groups:

Polymer Matrix Composites (PMCs) These are the most common and will be

discussed here as this work comes under this category. Also known as FRP - Fiber

Reinforced Polymers (or Plastics) these materials use a polymer-based resin as the

matrix, and a variety of fibers such as glass, carbon and aramid as the reinforcement.

Metal Matrix Composites (MMCs) - Increasingly found in the automotive industry,

these materials use a metal such as aluminium as the matrix, and reinforce it with fibres,

or particles, such as silicon carbide.

Ceramic Matrix Composites (CMCs) - Used in very high temperature environments,

these materials use a ceramic as the matrix and reinforce it with short fibres, or whiskers

such as those made from silicon carbide and boron nitride.


20/53

9

4.3 Polymer matrix composites

Resin systems such as epoxies and polyesters have limited use for the manufacture of

structures on their own, since their mechanical properties are not very high when

compared to, for example, most metals. However, they have desirable properties, most

notably their ability to be easily formed into complex shapes. Materials such as glass,

aramid and boron have extremely high tensile and compressive strength but in solid

form these properties are not readily apparent. This is due to thefact that when stressed,

random surface flaws will cause each material to crack and fail well below its theoretical

breaking point. To overcome this problem, the material is produced in fibre form, so

that, although the same number of random flaws will occur, they will be restricted to a

small number of fibres with the remainder exhibiting the materials theoretical strength.

Therefore a bundle of fibres will reflect more accurately the optimum performance of the

material. However, fibres alone can only exhibit tensile properties along the fibres length,

in the same way as fibres in a rope.

It is when the resin systems are combined with reinforcing fibres such as glass, carbon

and aramid, those exceptional properties can be obtained. The resin matrix spreads the

load applied to the composite between each of the individual fibres and also protects the

fibres from damage caused by abrasion and impact. High strengths and stiffnesses, ease

of moulding complex shapes, high environmental resistance all coupled with low

densities, make the resultant composite superior to metals for many applications. Since

PMCs combine a resin system and reinforcing fibres, the properties of the resulting

composite material will combine something of the properties of the resin on its own with

that of the fibres on their own.

Overall, the properties of the composite are determined by:

i) The properties of the fibre

ii) The properties of the resin

iii) The ratio of fibre to resin in the composite (Fibre Volume Fraction)

iv) The geometry and orientation of the fibres in the composite


21/53

10

The ratio of the fibre to resin derives largely from the manufacturing process used to

combine resin with fibre, as will be described in the section on manufacturing processes.

However, it is also influenced by the type of resin system used, and the form in which the

fibres are incorporated. In general, since the mechanical properties of fibres are much

higher than those of resins, the higher the fibre volume fraction the higher will be the

mechanical properties of the resultant composite. In practice there are limits to this, since

the fibres need to be fully coated in resin to be effective, and there will be an optimum

packing of the generally circular cross-section fibres. In addition, the manufacturing

process used to combine fibre with resin leads to varying amounts of imperfections and

air inclusions. Typically, with a common hand lay-up process as widely used in the boat-

building industry, a limit for FVF is approximately 30-40%. With the higher quality,

more sophisticated and precise processes used in the aerospace industry, FVFs

approaching 70% can be successfully obtained.

The geometry of the fibres in a composite is also important since fibres have their highest

mechanical properties along their lengths, rather than across their widths. This leads to

the highly anisotropic properties of composites, where, unlike metals, the mechanical

properties of the composite are likely to be very different when tested in different

directions. This means that it is very important when considering the use of composites to

understand at the design stage, both the magnitude and the direction of the applied loads.

When correctly accounted for, these anisotropic properties can be very advantageous

since it is only necessary to put material where loads will be applied, and thus redundant

material is avoided. It is also important to note that with metals the properties of the

materials are largely determined by the material supplier, and the person who fabricates

the materials into a finished structure can do almost nothing to change those in-built

properties. However, a composite material is formed at the same time as the structure is

itself being fabricated. This is a FUNDAMENTAL distinction of composite materials and

MUST always be considered during design and manufacturing stages.


22/53

11

4.4 Classical Lamination Theory of Composites

The purpose of this section is to go through the theory describing the behaviour of

laminated composites called classic laminate theory (CLT). This is the basis for the

software called ComposIT by Bureau Veritas, which is used in the later part of the report

in the calculation of the PLy and the laminate elastic properties. It is also used in the

synthesis, section 9, to adjust material properties so that the computed strains resemble

the experimentally found strains.

In classic lamination theory (CLT) the individual layers are treated as having material

values, both stiffness and strength independent of each other. The stiffness matrix (also

denoted Q) of the individual layer describes the relation between the strains in the layer

and the stresses in it.

Equation 1: Lamina Stiffness matrix for an orthotropic material. (Jones, 1999, p. 71)

Note that there are five independent values defined in Equation 1: E1, E2, G12, 12 and

21. According to (Jones, 1999, p. 72):

Thereby four independent in-plane stiffness properties (E1, E2, G12 and 12) are

necessary to define an orthotropic layer, like the UD layers in the mats. For isotropic

layers

Thereby two independent stiffness properties (E and v) are necessary to define an

isotropic layer like the CSM layers and polyester layers in the mats.

The stiffness matrix is defined in the fibre coordinate system: The 1-direction defines the

fibre direction and the 2-direction is the in-plane perpendicular direction to the first. This


23/53

12

can be seen in Figure 2. Therefore the stiffness matrix must be rotated to the laminate or

global direction according to Equation 2.

Layer (1-2) and Laminate/Global (x-y) coordinate systems

Equation 2: Transformation of stiffness matrix to global coordinates. Jones pp. 75-76, eq.

2.76, 2.77, 2.82

The global stiffness matrices are combined for all layers to a so-called ABD-matrix for

the plate. The ABD-matrix describes the plate stiffness, i.e. the relation between the

loading of the plate (See Figure 3) and the response in terms of strains and curvatures. In

fact the ABDmatrix is a gathering of three matrices (A, B and D):

Note that in the tests only Nx is applied to the specimens

A: Extension stiffness. Relation between force resultants (N) and in-plane strains ()

B: Bending extension coupling stiffness. Relation between force resultants (N) and

curvatures () and relation between moment resultants (M) and in-plane strains ().

D: Bending stiffness: Relation between moment resultants (M) and curvatures () of the

plate.

The ABD-matrix is calculated by adding the individual stiffness matrices of the layers

directed in the x-direction with the contribution of the thickness of the individual layers

as seen in Equation 3.


24/53

13

Figure 4.2 Loads on a plate

Equation 3: Calculation of A, B and D values. z is the coordinates in the thicknessdirection as shown in Figure 4.

Figure 4.3 z-coordinates of the individual layers through a laminate.


25/53

14

ABD-matrix of a laminate. (Jones, 1999, p. 244) is given by the above matrix.

4.5 Sandwich Structures

Single skin laminates, made from glass, carbon, aramid, or other fibers may be strong,

but they can lack stiffness due to their relatively low thickness. Traditionally the stiffnessof these panels has been increased by the addition of multiple frames and stiffeners,

adding weight and construction complexity.

A sandwich structure consists of two high strength skins separated by a core material.

Inserting a core into the laminate is a way of increasing its thickness without incurring

the weight penalty that comes from adding extra laminate layers. In effect the core acts

like the web in an I-beam, where the web provides the lightweight separatorbetween

the load-bearing flanges. In an I-beam the flanges carry the main tensile and compressive

loads and so the web can be relatively lightweight. Core materials in a sandwich structure

are similarly low in weight compared to the materials in the skin laminates.

Engineering theory shows that the flexural stiffness of any panel is proportional to the

cube of its thickness. The purpose of a core in a composite laminate is therefore to

increase the laminates stiffness by effectively thickening it with a low-density core

material. This can provide a dramatic increase in stiffness for very little additional

weight.

Figure shows a cored laminate under a bending load. Here, the sandwich laminate can be

likened to an I-beam, in which the laminate skins act as the I-beam flange, and the core

materials act as the beams shear web. In this mode of loading it can be seen that the


26/53

15

upper skin is put into compression, the lower skin into tension and the core into shear. It

therefore follows that one of the most important properties of a core is its shear strength

and stiffness.

Figure4.4 Sandwich Panel loading

In addition, particularly when using lightweight, thin laminate skins, the core must be

capable of taking a compressive loading without premature failure. This helps to prevent

the thin skins from wrinkling, and failing in a buckling mode.

4.5.1 Structural Design Criteria

Sandwich structures should be designed to meet the following structural criteria:

Skin facings should be sufficiently thick to withstand tensile, compressiv e and in-plane

shear stresses induced by the design loads.

The core should have sufficient strength to withstand transverse shear stresses induced

by the design loads.

The core should have sufficient flexural and shear stiffness to avoid excessive

deflections.

The core should be sufficiently thick and have sufficient shear stiffness to prevent panel

(or general) buckling of the sandwich under load.

Compressive modulus of the core and compressive strength of the facings should be

sufficient to prevent wrinkling of the faces under design loads. The core should have sufficient compressive strength to resist crushing by design loads

acting normal to the panel facings or flexure induced compressive stresses.

Material strength in the vicinity of cut -outs and attachments should be sufficient to

prevent failure in these regions of stress concentrations.


27/53

16

4.5.2 Facing/Skin Design

Under flexural (or bending) loads one face is in tension and the other is compression. If

the stresses in either facing exceed the corresponding ultimate stresses of the constituent

materials of the facings, then the sandwich panel will fail in a catastrophic manner. In

design calculations, the strength verification of the facings is usually carried out by

comparing the stresses caused by external loads with the allowable stresses for the

constituent materials of the facing. The allowable stresses are obtained by dividing the

strengths by suitable factors which take into account the variable properties of the

materials, the approximations in structural design, accidental loads, fatigue performance,

etc. When the calculated stresses exceed the allowable stresses, a change in sandwich

design is required.

Use a material with higher allowable stresses (i.e. strengths) for the facing;

Increase facing thickness, thus reducing the applied stresses; or

Increase core thickness, thus reducing the applied stresses (preferred method).

Note: A higher density (i.e. stiffer) core does not affect the stresses in the facings.

4.5.3 Core Design

If the shear stress induced in the core is greater than the shear strength of the core

material the core will fail, resulting in failure of the sandwich structure. As before, the

allowable shear stress is obtained by dividing the shear strength of the core by a suitable

safety factor. When the calculated shear stress exceeds the allowable shear stress, a

change in sandwich design is required. In such a case, the following is recommended

Use a core material with higher allowable shear stress; or

Increase the core thickness (preferred method).

Note: Using a different material for the skins or increasing skin thickness has no affect

on the shear stress in the core.


28/53

17

5. Material Selection

A sandwich can be correlated to a system where different subsystems interact to form a

stiff structure. Unlike a plate of steel or aluminium where there is only one isotropic layer,

here the layers of materials interact with each other at different levels depending on their

orientation and individual properties. The various parts of a sandwich are to be looked

into, the requirements of each part and their selection on the basis of fulfillment of those

requirements. This selection of materials is done by an iterative process, followed by the

calculations. It should be noted that at this stage of the project, the cost based

considerations are not done, as that involves contacting the individual supplier for

individual quotes.

5.1 Core

The requirements on the core of a sandwich say that it should have low density, high

compressive and shear strength, and high shear modulus and be thermally stable. It

should also posses a good resistance to water penetration and have good fatigue strength.

DIAB is a popular choice for cores and hence the calculations are carried out using

information from their product catalogues. The Divinycell structural foam cores offer

good mechanical properties; they are classified as H, HT, HP, HCP and HD grade. TheHCP core is strong and heavy and unsuitable for this purpose. Further the H, HP and HT

grade of cores are split into cores of different densities and strengths.

End grain Balsa (i.e. grains oriented in the through thickness direction) is an efficient

core, providing good strength with low density. The major problem with Balsa is

susceptibility to water penetration, leading to swelling, debonding and rotting. DIAB also

has DNV approval on the Balsa cores. They have mainly three kinds of Balsa cores, but

all three of them are heavy when compared to the available PVC cores. Hence only the H

series cores have been taken into consideration. All calculations have been performed

with the H100 core.


29/53

18

Property Unit H60 H100

Nominal Density Kg/m 60 100

Tensile Strength MPa 1.8 3.5

Tensile Modulus MPa 75 130

Shear Strength MPa 0.76 1.6

Shear Modulus MPa 20 35Table 5.1 Properties of different core

5.2 Laminates

A face or a laminate consists of a resin system and the fibres that are embedded into it. A

laminate comprises of plies or laminas that are built up of a fibres and resins. The fibre

that could be used for making the faces of structure should be light and strong.

5.2.1 Fibres

The fibres are responsible for taking a majority of tensile and compressive loads. An

initial investigation into the kind of fibres that can be used for this project resulted in E-

glass, S-glass, Carbon fibre and Aramid Kevlar. The poor compressive strength of

Aramid acted as a catalyst in disregarding it. The major advantage of E-glass is that it is a

high strength and low cost material. The high strength S glass has slightly better

mechanical properties than E-glass and is also more expensive. Cost considerations limit

the use of Carbon fibres. Some properties of the fibre materials in consideration are asfollows

Property Unit E-Glass

Youngs Modulus GPa 72.4

Poissons Ratio 0.2

Shear Modulus GPa 30

Mass/mof fiber g/m 600

Table 5.2 Properties of the Fiber

5.2.2 Resin System

The most commonly used resin in marine applications is polyester. It is moderately

priced and is easy to handle. Vinyl ester, Phenolics and Epoxies are also used depending

on the application. In case of a fire, a phenolic based laminate would burn on the outside


30/53

19

forming a char. This char prevents the burning of the inner material. Their initial physical

strength is slightly lower than the polyesters however, in the case of a fire; they are

capable of maintaining their strength for a longer time period and to a higher temperature.

But being very viscous, phenolics also exhibit issues when it comes to infusion. The

various physical properties of the resins studied are

Property Unit Polyester Epoxy

Density Kg/m3 1210 1200

Youngs Modulus GPa 3.6 3.0

Poissons Ratio - 0.36 0.37

Table 5.3 Properties of the Resin

5.2.3 Calculation of Ply and Laminate Properties

The properties of ply is calculated using rule of mixtures and this is done by ComposeIT

software by Bureau Veritas figure 5.1. For using in the theoretical calculation the

laminate properties are also computed in figure 5.2.

o

Figure 5.1 Calculation of Ply properties

Figure 5.2 Calculation of Laminate Properties


31/53

20

6. Manufacturing Process

The sandwich panels are made with Vacuum Assisted Resin transfer Molding. This is

superior to the conventional Hand layup technique.

Figure 6.1 Vacuum Assisted Resin transfer Molding

Large steel moulds are made with the shape of the covers. These moulds can be used

800-1000 times before they must be replaced. First the gel coat is applied on the mould

surface. Then the outer fibre mats, cut into size of the mould, are laid into the gel coat.

On top of that the plates of foam are placed and then the inner fibre mats. A layer of peel

ply is laid on top of the fibres with a net or a cloth mat above to soak excess resin. On top

of all, a vacuum bag is placed and sealed at the edges to create an airtight space around

the mats and foam. A pipe is inserted under the vacuum bag at each end of the mould.

One is connected to the polyester resin and the other to a vacuum pump. The vacuum

pump is started causing around 80% vacuum in the mould, thereby pulling the resin

through the mould imbedding the fibres and the foam. In addition it also removes most of

the air inclusions in the laminate. When all mats and plates are imbedded, the vacuum

pump is stopped. The cover hardens in the mould for some hours before it is taken out.


32/53

21

7. Design Considerations

Figure 7.1 Direction of Anchor Point Loading

1.

The roof of the nacelle cover is modular by nature and is connected to the metal

frame by a single row of bolts, [6] claims that such a connection needs to be

considered as simply supported condition.

2. The panel is considered as a beam in place of a shell

3. The effective youngs modulus of the laminate is calculated with the ComposIT

software, Bureau Veritas, by taking in to consideration the resins (polyester)

properties also.

4. A panel of length 2000 mm and breadth 600 mm is considered for the calculations

5. There are several load cases acting on the nacelle cover such as wind load, snow

load, personnel loading and so on, out of which only anchor point loading

conditions are considered

6. Thickness of the top and bottom facings are assumed to be same, t1= t2= tf

7. The structure is considered to be symmetric about the neutral axis.

8. The Lay-up chosen should be approximately quasi-isotropic (i.e. based on 0o,

45oand 90

oplies) [7],[8].

9.

The grouping of the 90oand the 0oply should be avoided [8]

10.The thickness of the facing is increased by increasing the number of quasi-

isotropic laminates

11.Two quasi isotropic layups are used on either side of the core. Hence that gives us

the stacking sequence as [90/-45/45/0/90/-45/45/0/CORE/0/45/-45/90/0/45/-45/90]


33/53

22

12.The curvature of the nacelle roof is not considered in the design

13.The Anchor Points should carry a load of 22.2 kN [10] for use in North America

and this load varies between countries. In this project the design is done for a load

of 36 kN.

7.1 Partial safety factor method

The limit state function can be separated into load and resistance functions S and R so

that the condition becomes [1]

The resistance R generally corresponds with the maximum allowable design values of

material resistance, hence R(fd) = fd, whilst the function S for ultimate strength analysisis usually defined as the highest value of the structural response, hence S(Fd)=Fd. The

equation then becomes

Fk- Design Load

fk- Resistance offered by material

f- Partial Safety factor for load = 1m - Partial safety factor for materials = 2.4

n- Partial safety factor for consequence of failure = 1


34/53

23

8. Theoretical Calculation

The procedure for calculation is iterative by nature; the steps adopted in the calculation

are as given below. Initially a set of values are assumed and then the output are compared

to the allowable limits. Based on the comparison increment or decrement is done and

final set of values are arrived upon. The following calculations show the final set of

values arrived upon. Initially a H60 core was considered and one quasi-isotropic plate of

1.868 mm was considered for the facing. Based on the system response to the increment

was done and the same are tabulated in the table 9.1. In order to validate the calculation a

simple stacking sequence is taken and calculated and compared with ANSYS. After the

validation the Anchor Point loading conditions are applied through multi point constraint

option.

I.

Define loading conditions

II. Define panel type

III. Define physical/space constraints

IV. Calculate

Figure 8.1 Simply supported Boundary Condition

Formulas

1. Bending Stiffness - D

2.

Shear StiffnessS

3. Net Deflection = Bending Deflection + Shear Deflection


35/53

24

4. Facing Stress

5. Core Stress

Input Units Nomenclature

Ef 1.386E+10 Pa Effective Youngs Modulus of the facing sheet

tf 0.003736 m Facing thickness

h 0.045 m Distance between facing skins = tf + tc

Gc35000000 Pa

Shear modulus of the core

Kb 0.0208333 Beam - bending deflection coefficient

SSB

Ks 0.25 Beam - shear deflection coefficient

SSB

P 36000 N Load acting on the Anchor Point

L 2 m Length of the beam

b 0.6 m Beam width

M 18000 Nm Maximum bending moment - (P*l)/4

SSB

F 18000 N Maximum shear force

SSB

tc 0.041264 m Core thickness

1 8.28E+8 Pa Maximum facing stress of facing material

1 1.6E+6 Pa Maximum core shear strength

Table 8.1 Input Parameters


36/53

25

Output

Bending stiffness - D 31456.933 Nm2

Shear stiffness - S 945000 N

Bending Deflection 0.190737 m 190.737 mm

+Shear Deflection 0.0190476 m 19.04762 mm

=

Net Deflection = Bending def + Shear

def 0.2097846 m 209.7846 mm

Facing Stress 178443969 Pa 178.444 Mpa

Core Stress 666666.67 Pa 0.666667 Mpa

Table 8.2 Output Parameters

Based on the partial safety factor method explained in the previous chapter, the constants

for the partial safety of various parameters are obtained from [1]. They are substituted in

the following equation and compared.

f Partial Safety factor for load 1

m Partial safety factor for materials - composites 2.4

n Partial safety factor for consequence of failure 1

Table 8.3 Partial safety factors

1. Check for facing Stress

Units Design Stress Permitted stress Compare & Decide

Mpa 178.444 345 Safe

2. Check for core shear stress

Mpa 0.666667 0.6666667 Safe

Table 8.4 Check for design loads versus material's resistance


37/53

26

8.1 Validation Of The Theoretical Model

For the sake of validation, a simple sandwich structure is taken with the core between two

unidirectional plies on top and bottom [0/CORE/0]. This validation is mandatory to

determine, how good the theoretical model depicts the actual system.

Input Units Converted Input Units Nomenclature

Ef 38.325 GPa Ef 3.83E+10 Pa

Effective Youngs Modulus of the

facing sheet

tf 0.467 mm tf 0.000467 m Facing thickness

h 40 mm h 0.04 m Distance between facing skins = tf + tc

Gc 40 MPa Gc 40000000 Pa Shear modulus of the core

Kb 0.020833 Kb 0.020833 Beam - bending deflection coefficientSSB

Ks 0.25 Ks 0.25 Beam - shear deflection coefficient

SSB

P 3.6 KN P 3600 N Load acting on the Anchor Point

L 2 m L 2 m Length of the beam

b 0.6 m b 0.6 m Beam width

M 1800 Nm Maximum bending moment

F 1800 N Maximum shear force

tc 39.533 mm tc 0.039533 m Core thickness

Bending stiffness - D 8590.932 Nm2

Shear stiffness - S 960000 N

Bending Deflection 0.069841 m 69.84108 mm

+

Shear Deflection 0.001875 m 1.875 mm

=Net Deflection = Bending deflection

+ Shear deflection 0.071716 m 71.71608 mm

Facing Stress 1.61E+08 Pa 160.5996 Mpa

Core Stress 75000 Pa 0.075 Mpa

Table 8.5 Validation theoretical model


38/53

27

14.FEA Validation

9.1 Preprocessing Model Validation

The dimensions of the panel 2m X 0.6mThe edges are simply supported

Load applied 3600 N

Meshing Element edge length 0.025 m

SHELL91 is chosen - The core is assumed to carry all of the transverse shear,

while the faceplates carry none; conversely, the faceplates are assumed to carry

all (or almost all) of the bending load. Only SHELL91 has this sandwich option.

No of layers = 3 [0/CORE/0]

Keyopt K6 = 1

Keyopt K8 = 1

Keyopt K9 = 1

Keyopt K11 = 1

Under Real constants put Tk = thickness of the ply = 0.467mm

Material property of the core is given for H100 as given in matweb website

Unidirectional Ply properties is given from the values computed by the

ComposeIT software


39/53

28

Figure 9.1 Meshing and Boundary conditions

9.2 Post Processing - Model Validation

The overall deflection of the system is given by the displacement vector sum

The facing stress and the shear stress are also plotted and matched with the

theoretical results


40/53

29

Figure 9.2 Deflection observed

Figure 9.3 Facing Stress


41/53

30

Figure 9.4 Core Shear Stress without edge effects of Layer 2

Comparing these ANSYS values to the theoretical model validates the theoretical model

Output Theoretical ANSYS%Variation

Net Deflection = Bending

def + Shear def 0.071716 m 71.71608 mm 67.1 mm 6.436609

Facing Stress 1.61E+08 Pa 160.5996 Mpa 152 mpa 5.354667

Core Stress 75000 Pa 0.075 Mpa 0.07774 Mpa -3.65333

Table 9.1 Comparison between theoretical and actual values

Hence we infer that the theoretical model represents the system in consideration. Hence

using this model, a panel could be designed which can take the said 36kN load. The

Anchor Points sets up a load on to the system in a unique manner. Hence the system is

that is designed for the point loading condition is tested for anchor point loading

condition and allowable limits are checked.


42/53

31

9.3 Preprocessing Multi Point Constraint

In order to simulate the anchor point loading conditions, the exact assembly of the anchor

points with composite covers has to be done and proper contact constraints has to be set

between the bolt and the composite cover.

Figure 9.5 Anchor Point Loading

However, since the system in focus is the composite cover, similar boundary conditions

are simulated using multi point constraints at 51 mm above the sandwich panel. Exactly

at the centre of the panel at a height of 51mm, a pilot node is created and the nodes in the

base of the anchor points are attached to that pilot node. When a force is setup on the

pilot node, it is transferred to the nodes below. A force of 36kN is set up along the x

direction in the pilot node.

The dimensions of the panel 2m X 0.6m

The edges are simply supported

Load applied 3600 N

Meshing Element edge length 0.025 m

SHELL91 is chosen - The core is assumed to carry all of the transverse shear,

while the faceplates carry none; conversely, the faceplates are assumed to carry

all (or almost all) of the bending load. Only SHELL91 has this sandwich option.

No of layers = 3 [0/CORE/0]

Keyopt K6 = 1


43/53

32

Keyopt K8 = 1

Keyopt K9 = 1

Keyopt K11 = 1

Under Real constants put Tk = thickness of the ply = 0.467mm

Material property of the core is given for H100 from the website matweb

Unidirectional Ply properties is given from the values computed by the

ComposeIT

Figure 9.6 MPC-Meshing and Loading conditions


44/53

33

9.4 Post Processing Multi Point Constraint

Figure 9.7 Equivalent Stress in layer 1



45/53

34




46/53

35

Figure 9.11 Stresses along the X direction

Figure 9.12 Shear Stress in layer 9 Core


47/53

36

9.5 Results and Discussion

The value of the facing stress and the core shear stresses are well with in the permitted

limits. In the theoretical calculations the allowable stress are computed with m = 2.4. That

was without considering stress concentration caused by the holes. Since the analysis done

here is specifically for the Anchor Point Loading conditions, the stress concentration

needs to be considered and according to [1] m has to be incremented by 2 in this regard.

Hence the new value becomes m= 4.4. Hence the new allowable limits are Facing Stress

= 188.181 MPa and the core shear stress = 0.363 MPa.

From figures 9.8, 9.9, 9.10, 9.11 it can be inferred that the facing stress of the system was

found to be 20.2 MPa compared to the allowed 181.181 MPa

By observing the plots it can be inferred that the resistance offered by the system is much

uniformly distributed about the base of the anchor point in the plate and it does not get

transferred to the edges. Hence the load acting on the system is locally transferred in an

effective way and the design is safe from the facing materialsperspective.

From figure 9.13, it can be inferred that the shear stress is 0.074 MPa which is well below

the permitted limit of 0.363 MPa and hence it can be said that the core is safe.

Since the facing stress and the core shear stress are within limits, it can be said that the

said sandwich structure is qualified. Optimization of this design has to be done before

implementing in the actual nacelle.


48/53

37

10. Conclusion

1. Thus the theoretical calculation for the sandwich panel was done for the given

loading conditions.

2. The model was validated using ANSYS with a variation of 10% between the

theoretical and the actual.

3. To simulate the anchor point loading condition, multi point constraints was used and

a force of 36000 N was applied to a pilot node that is connected to all the nodes of the

anchor points base area.

4. The loading that is applied on to the composite cover is distributed in a uniform

manner, which can be seen from the plots.

5. The observed stresses caused by the load are well within the allowable limits thereby

qualifying the design.


49/53

38

11. Reference

1. Standards: IEC 61400-1 Standard for Wind Turbines; 3rd edition, 2005;DS/ENISO 527-4

2.

Determination of tensile properties part 4: Test conditions for isotropic andorthotropic fibre-reinforced plastic composites; 1. Edition, 1996; DS/EN ISO14126

3. Determination of compressive properties in the in-plane direction Fibre reinforced

plastic composites 1. Edition 1999

4. Jones, Robert M.; Mechanics of Composite Materials,2nd edition, 1999;ISBN 1-56032-712-X

5. Kreyszig, Erwin, Advanced Engineering Mathematics, 8th Edition, 1999; ISBN

0-471-33328-X

6. "Plastics Products Design Handbook" by Mr. Marcel Dekker, McGill Corporation

7. Composites Engineering Handbook, Editor Mallick, P.K., Marcel Dekker,

1997.

8. Military Handbook, Polymer Matrix Composites, Volume 3 Materials Usage,Design and Analysis, MIL-HDBK-17-1E, 1994.

9. EN795/1995 - EN795/A1:1999 Protection against fall from heights-anchor

devicesrequirements and testing

10. ANSI/ASSE Z359.1-1992(R1999) Safety Requirements for Personal Fall ArrestSystems

11.EN 795: 1996 Protection against falls from heights Anchor devices

Requirements and Testing

12.EN 50308: 2004 Wind turbineProtective measures Requirements for design,

operation and maintenance

13.

B.T. Astrom, Manufacturing of Polymer Composites, Chapman & Hall, 199214.Issac M Daniel and Ori Ishai, Engineering mechanics of composite materials;

Oxford Publishers; 2006

15.Bruhn EF. Analysis and design of flight vehicle structures. Purdue University,

West Lafayette, IN: S.R. Jacobs and Associates, Inc.; 1973. p. C12.1C12.52


50/53

39

16.Hexcel Corporation. Design Handbook for Honeycomb Sandwich Structures.Technical Service Bulletin #123. Huntington Beach, CA: Hexcel Corporation;

1970.

17.Vinson JR. The behaviour of sandwich structures of isotropic and composite

materials. Westport: Technomic; 1999.

18.Mines RAW, Worrall CM, Gibson AG. Low velocity perforation behaviour of

polymer composite sandwich panels. Int J Impact Engng 1998;21(10):85579.

19.Torre L, Kenny JM. Impact testing and simulation of composite sandwichstructures for civil transportation. Compos Struct 2000;50:25767.

20.Kootsookos A, Burchill PJ. The effect of the degree of cure on the corrosion

resistance of vinyl ester/glass fibre composites. Composites A 2004;35:5018.

21.

Allard JF. Propagation of sound in porous media: modelling sound absorbingmaterials. Elsevier Applied Science; 1993.

22.Sahraoui S, Mariez E, Etchessahar M. Mechanical testing of polymeric foams at

low frequency. Polym Test 2001;20:936.

23.Anderson, Melvin S. Optimum proportions of truss core and webcore sandwichplates loaded in compression. NASA TN D-98; 1959.

24.Steeves CA, Fleck NA. Material selection in sandwich beam construction. Scripta

Mater 2004;50:13359.

25.

Hexcel Corporation. Design Handbook for Honeycomb Sandwich Structures.Technical Service Bulletin #123. Huntington Beach, CA: Hexcel Corporation;

1970.


51/53

40

Appendix A


52/53

41

Appendix B


53/53

Documents

Design and Analysis of a GFRP Sandwich Panel Subjected to Anchor Point Loading Conditions