Accepted Manuscript

Design of Composite Tank Covers

Agnieszka Bondyra, Marian Klasztorny, Aleksander Muc

PII: S0263-8223(15)00550-4
DOI: http://dx.doi.org/10.1016/j.compstruct.2015.07.008
Reference: COST 6585

To appear in: Composite Structures

Please cite this article as: Bondyra, A., Klasztorny, M., Muc, A., Design of Composite Tank Covers, Composite
Structures (2015), doi: http://dx.doi.org/10.1016/j.compstruct.2015.07.008

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
 Design of Composite Tank Covers
Agnieszka Bondyraa*, Marian Klasztornyb, Aleksander Muca
a
Department of Machine Design, Faculty of Mechanical Engineering, Krakow University of
Technology, Al. Jana Pawla II 37, PL-31864 Krakow, Poland, e-mail: abondyra@wp.pl,
olekmuc@mech.pk.edu.pl
b
Department of Mechanics and Applied Computer Science, Faculty of Mechanical
Engineering, Military University of Technology, ul. gen. S. Kaliskiego 2, PL-00908 Warsaw,
Poland, e-mail: m.klasztorny@gmail.com
*
Corresponding author (phone: +4812 3743409)

Abstract. The problem of design and failure analysis of composite tank covers is discussed
in the paper. Covering systems consist of segmental parts of tanks, independently on tank
structure. Composite covers have several advantages connected with structure, transport and
assembly. Additionally, the construction divided into several sections enables repair and
inspection improving. However, on the one hand, failure of the tank covers, which are
occurred sometimes, and on the other hand new applications, make it necessary to develop
design, analysis and optimization of these structures. In this study, the main advantages of the
GRP covers are shown against other covers and roofs systems. The construction of the
representative rectangular composite tank cover, segments and joints are described. Failure
analysis of the covering segment under combined loading is presented. The exemplary cover
was built in the sewage treatment plant in Poland. The cover is built of repetitive two-wave
segments with flanges overlapping one another. The segments are made of mixed glass-fibre
reinforced plastic laminates. The cover is subjected to the following static loads: dead load,
technological vacuum, snow and ice and local load (workers with tools). Static analysis of
the exemplary cover was conducted using the finite element code MSC.Marc. The
considerations include shell geometry of the cover, structural GRP laminates with specified
ply sequences, segment connections of rivet nut–bolt and anchor–nut types, friction at the
interface between flanges and between flanges and tank walls. Safety coefficients have been
analyzed for the different elements of the exemplary composite cover.

Keywords: composite covers, composite joints, finite element method, static analysis, failure

1. Introduction

Roofs and covering systems are well-known solutions in typical building constructions
such as houses, gyms, stores etc. There is a wide range of original applications to cover the
engineering construction, e.g. on the one hand steel and steel-concrete composite shell roofs
and on the other hand the composite covering systems. Self-supported metallic roof shells can
perform a double function: the arch beam and the building cover [1, 2]. Another interesting
solution design, steel–concrete composite shell roofs (Comshell roofs), are formed by pouring
concrete on a thin stiffened steel base shell (the permanent formwork with the tensile steel
reinforcement). It is constructed by bolting together open-topped modular units consisting of
a base plate with surrounding edge plates to form with thin stiffeners in both directions [3].
Concrete [4, 5], steel [6, 7] tank dome roofs are widely described in literature. Concrete tanks
constructions have sometimes construction errors caused by the location of the reinforcement
layers, which can be discovered as a consequence of the failure. Natural hazards are possible
to additionally hasten construction collapse [5].

1
 The shape of above-ground tank covers is one of the classify criterion of the tank types,
due to the fact that general dimensions of the tanks and covers result from the design. There
are several standardized steel tank roofs: fixed (domes or spherical caps), conical (a vertical
axis of symmetry, the flat bottom, shallow cone top), umbrella (self-supporting structure,
partial construction), dome (flatter than umbrella). A special type of a tank roof is a floating
cover, which has a form of an external or internal structure in order to reduce the area of a
liquid, that is exposed to the atmosphere influence. Design requirements for steel tank roofs
are included in calculation under combinations of loading: fluid and internal pressure,
hydrostatic test, wind and internal pressure, gravity and seismic loads [8].
Steel or concrete roofs are used to cover silos, liquid tanks, warehouses, exhibition halls
etc. They are made of segments, rings, cover plates and additional elements. Finite element
analysis (FEA) allowed improvement of existing roof systems, design of such structures and
analysis of these systems under various loading combination. Methodologies of covering
systems calculations presented in the literature demonstrate the necessity of taking into
account joints between composite segments [9].
There is a variety of covering systems applications: for tanks with and without a central
support or walkway, for different tanks shapes, etc. Currently, this type of constructions has
been widely used in the applications such as water reservoirs, sewage treatment plants, above-
ground tanks for liquids, channels, spiral pump stations, special design tanks [10]. These
constructions required the additional features such as, e.g. encapsulation or lightweight, what
results in an increased use of GRP (Glass Reinforced Plastics). Development of composite
materials applications allowed design of modern tank covers. GRP composites and their
advantages were well known: high tensile and compressive strength, low weight, chemical
resistance, good color ability and UV resistance, maximum options for shaping and also
extreme resistance to weathering and ageing – GRP covers, which were assembled more than
thirty years ago, are still in use. The main disadvantage of composites, used for covers
constructions, is the low transverse shear strength of composites (delamination) [11].
Manufacturers [10] make covering systems from glass fiber components in the form of
mats, fabrics and non-woven, with strict adherence to the structural analysis and with
appropriate safety margins. Stainless steel is used for joints and additional elements. Cut
edges and drilled holes are carefully sealed for dependable prevention of moisture penetration
into the laminate. Structural analyses are carried out for all of the variants of their modular
system, with special consideration into the dimensions relating to the specific case and
perhaps specific loads, such as high wind or snow loads [11].
There are two main approaches in the composite covers design: a column beam and in-fill
stables arrangement (rectangular and small cylindrical tanks) and self-support construction,
built from modular sections in a conical or domed shape (large circular tanks).
The laminated composite covers are preferred to engineers, in comparison to steel and
concrete, constructions due to several advantages, such as:
- light weight,
- corrosion resistance,
- access and removal of sections of the covers,
- architectural consideration,
- cost effectiveness and flexibility in design (simplicity of modification),
- reliable structural analysis,
- inexpensive production precisely matching the specification,
- easy transport and assembly,
- offer flexibility to optimize the stiffness and strength properties.

2
 Industrial structures present a large variety of geometries, stacking sequences, constraints
and types of loadings. Such variety requires adequate meshing techniques, types of analysis
and material models to be adopted at the different stages of design and development.
For composites structures, there is an additional need to account for failure and damage, as
these may initiate at relatively low stress levels. It is therefore important to understand how
damage modeling approaches are affected by the many factors listed above.
The present contribution focuses on the modeling techniques for failure in structures made
of unidirectional glass reinforced polymers.
Static calculation and design methods of a rectangular tank composite cover are in the
early stages of research and development [10, 12, 13], however, in the literature, another types
of composite tank covering cases are described in more details.
Composite covers are made of segments, for example in a fluted shape. In tanks with a
smaller diameter, it is possible to minimize costs of constructions by using flat sandwich
panels. The considered, in the literature, circular covers were under loading by: live load
(loaded in the small area of about 1/3 up to the whole length of a spans), body load (gravity),
central load (in the central of cover) and thermal load (60°C, which is ignored in analysis
because of a small influence to deflection). All factors are applied as close to natural hazard in
order to get realistic results. The simple models with using the symmetry planes and limited to
one or several segments are shown in literature. Bolt spacing is also under considerations.
Design constraints are limited to maximum deflection of 30 mm [14].
The covers are mainly constructed in order to reduce the contact of the liquid in the tank
with the atmosphere. Segments are jointed with bolts. Polyurethane is used to seal the flanges
connections. Clearances in holes are designed to allow thermal displacements of composite
segments [15].
Estimation of the critical loading, causing covers damages, should also consider the
interaction between segments of the covering system to obtain proper values, such in the steel
roofs [16, 17]. However, full specification of composite tanks covers design includes several
main steps:
1. Definition of the tank shape, the tolerances, the stated dimensions
2. Concentration of the additional equipment such as bridges, pipework etc.
3. Definition access hatches for inspection/cleaning etc.
4. Design loading conditions: wind, snow/sand, dead loading (staff): occasional access
(cleaning), frequent access (inspection) and areas and walkways.
In the present work, rectangular composite covers are considered, whereas additional
equipment and access hatches are omitted. Design rules for loading conditions of composite
tank covers are discussed and verified with the use finite element method (FEM).

2. Composite cover joints modeling

Joints between segments with flanges and between whole covers and a tank wall (concrete)
are an important aspect of the covers assembly. In the literature review, mainly bolt joints
were used to connect the covers parts [9, 10, 14, 15, 18]. However, tank covers segments are
sometimes connected with each other by a rivet nut-bolt single lap joint and an anchor-nut
joint with the construction clearance between segments. In order to encapsulate the tank,
rubber gaskets are frequently placed in rivet nut joints. The cover is connected to the walls of
tanks with anchor-nut joints with the construction clearance. Experimental and numerical
results in reference to joints are carried out in Refs. [19, 20].
Classic bolt joints of laminates were analyzed by several authors. McCarthy et al. [21]
developed a 3D model of single bolt joints at the elastic range. A validated model from Ref.
[21] was used in Ref. [22] to test an influence of a micro clearance hole with a nominal

3
diameter of 8 mm on the distribution of stress and strain, rigidity of joints and initiation of
damage. The model of the bolt joints of composites was developed by Gray and McCarthy
[23]. Experiments and simulations are related only to the elastic range. Analytical approach to
modeling composites load distribution on the combined number of screw connectors have
been developed in Ref. [24], where joints and laminates are represented by a group of springs
and masses.
Macro-mechanical models of composites are formulated in accordance with a respective
lamination theory. In the case of 2D models of layered shells, equivalent single-layer theories
and discrete-layer theories are presented in the literature. This model used in conjunction with
the classical theory of thin shells of Kirchhoff-Love's is commonly known as the classical
laminate theory (CLT). Modeling of the elastic behavior of laminates with good accuracy
requires consideration of transverse shear deformation. These models for plates [25] and for
shells [26] require an appropriate transverse shear deformation factor due to constant shear
distribution across the shell thickness. In discrete layer theories (DLT), each lamina is taken
into account separately [27].
Strengths of composites involves determination of the stress limits of alternative complex
stress, carried by the material, which causes destruction of the material. Instead of limiting the
stress, strain can be considered as replacement collapsed deformation of the material, which
causes the material failure. So far, a number of strength hypotheses for composite layers have
been formulated [19, 28, 29].
Numerical modeling of laminated composite shells have been developed in several Refs.,
e.g. [13]. During the research, there were carried out numerical and experimental tests on
beam specimens in order to develop the options of the modeling and simulation of
homogeneous or mixed laminate up to failure, using the MSC.Marc software.

3. Analysis of composite tank covers

The design rules for composite tank covers are partly based of steel and concrete roofs,
but simplification of their static calculations frequently results in construction failure [15]. In
the absence of standards relating to composite covers of tanks, design standards of composite
pressure vessels and pipes standards are also used [30-35].

3.1. Description of the case study

The work considers is mainly devoted to the analysis of selected composite cover of a
rectangular underground tank. The cover is composed of two-wave segments with flanges
overlapping each other, arranged transversely in respect to the covering. The segments ensure
a slight decrease in bilateral longitudinal drainage.
Many repeatable parts allow a rapid manufacturing system. Additionally, these structure
can be optimized. The cover segments are made of mixed glass-fibre reinforced laminates
(layers with fabric, mat or roving reinforcement). The analyzed cover is under the following
static loads: dead load, technological vacuum, snow, ice and live load.
The study presents the case of a composite covering of Oxygen Stability Chambers Waste
Water Treatment Plant in Goleniow, hereinafter referred to as covering of KST [18]. The tank
is completely underground, a tank cover has a rectangular horizontal cross-section. Composite
cover is divided, by the work platform, into two parts, with dimensions in the plan ≈ 22 m ×
16 m each. In each section, there is a system of seven segments of a theoretical span
of l≈10 m. The details of the cover (sequences of layers, arrangement of the segments and
the other) are shown in Fig. 1-2 and Tables 1-3. The following notations are used: 1W – one-
wave segment, 2W – two-wave segment, TS - top segment, BS –bottom segment, R-N (rivet-

4
nut) joint – joint between two flanges, A-N1 (anchor- nut) joint – joint between the panel
flange and concrete wall, A-N2 (anchor nut) - joint between two flanges and concrete wall.
Bottom and top segments are arranged alternately (Fig. 1), i.e., each top segment interacts
with two bottom segments, and each bottom segment with two top segments. In the case of
the bisymmetrical load and bisymmetrical segments, there are two symmetry planes:
transverse vertical plane passing through the means of the span segments and longitudinal
vertical plane passing through the longitudinal axis segments. The example of cross-section
(A-A) were determined in order to panel slope shows. A computational subsystem is limited
to: 1/4 of TS and BS segments, 10 R-N joints, 3 A-N1 joints, 1 A-N2 joint.

Fig. 1. Goleniow cover computational sub-system [18]

Based on the design assumptions for the chosen composite covering, polymer matrix,
reinforcement, manufacturing technology, additives in materials, construction schedule
covering, the geometry of the segments and joints were selected. The cover segments were
manufactured in the hand layup method (the mat reinforced layers and fabric reinforced
layers) and spray technology (the cut glass fiber roving layers). The matrix resin was
polyester resin (Polimal 104 T). Protective layers (gelcoat, topcoat) have thickness of 0.35
mm. The sequence and thickness of the layers are set up in Table 1, where the following
notation is introduced:
STR600 – two-directional balanced plain weave fabric (600 g/m2) and stacking
sequence [0/90] with respect to the principal axes,
CSM450 –E-glass mat (450 g/m2),
CGR –spray cut E-glass fiber roving.

Tab. 1. The stacking sequence (top) of covering system KST [mm] [3]

Zone B Zone A
Flange
(0 - 0,25L) (0,25 l - 0,50 L)
CGR 1.30 2CSM450 1.80 5CSM45 4.50
STR600 0.95 CGR 1.30 CGR0 1.30
CGR 1.30 STR600 0.95 STR 600 0.95
STR600 0.95 CGR 1.30 CGR 1.30
STR600 0.95 STR600 0.95
total 4.50 total 6.30 total 9.00

5
 The experimental tests of the mechanical properties of laminates were performed
according to standards [18, 36-38] at temperature 20°C. Missing material constants were
determined from the conditions imposed on balanced orthotropic materials in plane 1, 2
[19, 36, 37], where eit , eic – ultimate normal strains at tension and compression in principal
directions ij = 12,23,31 , – ultimate shear strains in principal planes ij, ij = 12,23,31 and t, c are
indices denote tension and compression properties, respectively. Tables 2a and 2b summarize
the laminate material properties (a linear model of elastic –brittle material).

Tab. 2. a) Elastic constants [5]

Constant E1t E1c E3t E3c G12t G31

 [-] 23 [-] 31 [-]
[GPa] [GPa] [GPa] [GPa] 12 [GPa] [GPa]
STR600 16.5 18.9 17.7 5.00 0.16 0.76 0.23 2.28 2.40
CSM450 8.25 9.17 8.71 4.15 0.39 0.47 0.24 3.20 3.10
CGR 7.92 8.36 8.14 5.56 0.39 0.30 0.21 2.43 1.93

Tab. 2. b) Strengths and ultimate strains [5]

Strength R1t R3t R1c R3c S12 S31

[MPa] [MPa] [MPa] [MPa] [MPa] [MPa]
STR600 269 70 202 344 32.6 22.5
CSM450 95.7 70 216 231 91.1 35.9
CGR 72.3 70 170 177 87.7 42.3
Ultimate e1t e3t e1c e3c g12 g31
strain [-] [-] [-] [-] [-] [-]
STR600 0.021 0.020 0.011 0.102 0.05 0.045
CSM450 0.014 0.017 0.031 0.061 0.043 0.040
CGR 0.011 0.015 0.052 0.044 0.05 0.067

In the standard [32], there are three models of the design: 1st basic, 2nd basic and
advanced. In the case of Goleniow cover, 1st basic design conditions are fulfilled (the
material properties were determined on minimum of 5 samples). Partial safety coefficients are
assumed as follows [32]:

A1=1.3 (material tests),

A2=1.2 (chemical resistance),
T  20 50  20
A3  1  0.4 d  1  1.24 (influence of temperature) , (1)
HDT  40 90  40
A4=1.0 (influence of number of load cycles),
A5=1.7 (durability of covering - 30 years).
where: Td– design temperature, HDT – deflection temperature.
The global safety coefficient is calculated from standard equation:
n p  2 A1 A2 A3 A4 A5  2  1.3  1.2  1.24  1.0  1.7  5.0  nmin  3 . (2)

3.2. Composite joints

The segments are connected with the use of rivet nut-bolt joints and anchor-nut joints (Fig. 2).
Composite segments are joined together with R-B joints (rivet nut Ø 13 – bolt Ø 10), as
shown in Table 3.

6
 Fig. 2. Details of A-N1, A-N2, RN-B joints used in composite covering

Tab. 3. Segment characteristics of ½ covering system KST

width
type location comments
[mm]
2W 2180 BS technological holes, longitudinal support tank wall
1W 1090 TS holes Ø22 provide clearance for assembly of covering
2W 2100 BS holes Ø13 at the longitudinal edges of the assembly R-B
2W 2100 TS holes Ø22 provide clearance for assembly of covering
2W 2100 BS holes Ø13 at the longitudinal edges of the assembly R-B
2W 2100 TS holes Ø22 provide clearance for assembly of covering
2W 2100 BS holes Ø13 at the longitudinal edges of the assembly R-B
2W 2180 TS technological holes, longitudinal support tank wall

Encapsulation is applied by means of EPDM gaskets with dimensions b×h = 20×10 mm.
The connection of segments with concrete is realized using A-N1, A-N2 joints (anchor Ø 15 –
nut Ø 12). Boundary composite segments have built technological holes. Rivet nuts and
anchors are made of stainless steel A2-70.
In 2D modeling of the composite covering, two types of finite elements [39] were used: 75
– for laminate plates, 195 – for the RN-B joint. Element 75 (Bilinear Thick-shell Element) is
a four-node element with bilinear interpolation for the coordinates, displacements and
the rotations. This element can be used for curved shells in nonlinear analysis. This element
allowed to the interpolation of transverse shear strains to Gauss points after calculations
at the middle of edges. Element 195 (3-D Generalized Spring – CBUSH/CFAST Element) is
an elastic-damping structural element. Nonlinear stiffness properties can be defined directly
by defining three translational values and three rotational values or by a force-displacement
relation.
Element 195 has the assigned elastic properties of the connector (Bushing), translational
stiffness declared as wave forms change as a function of displacement forces in the global
coordinate system. Translational stiffness factors in x and y directions reflect the results
obtained for respective 3D RN B joint modelling. In the range of 0÷6 mm is free relative
sliding plate laminate, and after crossing 6 mm character of the joint stiffness is almost
constant. The stiffness characteristics (Fig. 3), which are necessary to be used in this type of
the finite element, were determined with numerical models of the tension tests.

7
 Joint Fxy1 Fz1 Fz2
type [kN] [MN] [MN]
A-N1 87.7 17.1 0
A-N2 79.2 13.9 0
RN-B 13.9 17.1 17.1

Fig. 3. Stiffness characteristics of joints

3.3. Loading combinations

The cover is assumed to be subjected to the following static loads: dead load (constant load
g), snow and ice load (variable load s), vacuum load (technological load p), workers with
tools (load q). Loads of g, p, q, s are varied linearly in the time of 1 s (g – gravitation, p –
underpressure, q – worker, s – snow). The load is considered in three variants:
1) subsequent loads g+p+q+s,
2) subsequent loads g+p+s+q,
3) simultaneous loads g, p, q, s.

Snow and wind load

Figure 4 shows the static scheme of the covering sub-system under snow loading. Snow
and ice load is calculated according to the formulas [12]:

s1k = 0.80  0.9  1.20  1 = 0.86kPa,

(3)
s 2 k = 1.60  0.9  1.20  1 = 1.72kPa.

Fig. 4. Static scheme of sub-system of Goleniow covering for snow loading

Snow and ice load is declared as a vertical component per unit area of the cylindrical shell,
which requires the conversion of the load from the horizontal to the oblique direction.

Vacum load
Vacuum load (within the tank) is declared by applying the radial pressure of 200 Pa
perpendicular to the surfaces of all finite elements of internal surfaces of the sub-system.

8
 Local load
Local load (Fig. 5) (a worker with tools) qk = 1.80 / 0.04 = 45kPa is declared as
the vertical loading applied to the surface with dimensions of 200×100 mm in the central part
of the sub-system of both the top and bottom segments.

Fig. 5. Deformation caused by local load (scale 10:1)

3.4. Analysis of composite tank cover

The numerical model

Scheme of the KST covering sub-system is built of the quadrant segments TS, BS,
concrete wall of the tank, and three symmetry planes. The model is consisted of 39 529 shells
elements. The concrete wall (to which the supporting flange segments are attached) is
modeled as a rigid surface, rounded off with a radius of 5 mm at the edge.
For modeling composite covering panels, shells elements with the dimensions~ 20× 20
mm were used. Main failure criteria, which can be used for composite panels, are available in
the MSC. Marc. The failure criterions, which describe the behavior of composite covering
system according to experimental tests, have been compiled in Appendix. An individual
lamina is modeled as a linear elastic–brittle orthotropic material. Hashin-Fabric failure
criterion for fabric-reinforced laminas and Maximum Strain failure criterion for mat / roving
reinforced laminas are selected.
For progressive composite failure modeling, Selective Gradual Degradation model is
selected. When failure occurs, the moduli are decreased gradually. When failure index F,
according to selected criterion, is higher than one stiffness reduction factors are calculated.
Residual stiffness factor, which describes the lowest values of stiffness, is adopted as 0.005.
Contact bodies are modeled as deformable-to-deformable system. Node-to-Segment
touching was selected to deformable composite panels. In the normal contact algorithm the
value of regular distance tolerance is selected as 0.15, bias factor is selected 0.95. A bilinear
model of the Coulomb friction with friction force tolerance 0.05 is used; other parameters are

9
obtained experimentally (Tab. 4). The bilinear Coloumb model assumes that the stick and slip
conditions are consistent with elastic and plastic relative displacements.

Tab. 4. Bilinear Coulomb friction model parameters

Friction couple Static friction cofefficient  [-] Ultimate stress gr [MPa]
composite - composite 0.39 91
composite - concrete 0.25 26.2

A full iterative Newton-Raphson procedure and the displacement and residual convergence
criteria with a tolerance of 0.05 are selected. The loading process is divided into 1000
increments (incremental steps). Full integration of layers is used. Failure indices of individual
lamina are applied to assess effort of laminate shells, defined according to selected failure
criteria. The safety coefficients of the covering system were determined from the load
condition [28]. Changes of default parameters of variable time-step incremental load of
minimum and maximum fraction of load case time from 1e-005 and 0.5 to 1e-004 and 0.001
have allowed achievement of better agreement between the experimental and numerical
results. In the work, the finite element analysis and the failure hypothesis were used to verify
failure mechanisms of the composite tank.
Figure 6 shows vertical displacement contours, due to all applied loads. Maximum values
of displacement z are equal 75.46 mm for bottom segment and 79.04 mm for top segement.
The effect of the interaction of segments TS, BS in transferring loads can be observed. It is
assumed that the segments cannot buckle.

Fig. 6. Vertical displacements contours [mm] and deformation of segments caused by full loading

Figure 7 illustrates the vertical displacement of central nodes in the BS, TS mid-span
sections as the time function (a linear increase of load). According to the theoretical
prediction, vertical deflection of the upper segment is a few percent higher than the deflection

10
of the bottom section. Influences of physical and geometric nonlinearities are minor. The
sequence of the load has a negligible effect on the final results.

80
g+p+q+s_BS
70 g+p+q+s_TS
g+p+s+q_BS
g+p+s+q_TS
60
g p s q_BS
g p s q_TS
50
w [mm]

40

30

20

10

0
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0
t [s]

Fig. 7. Change path of central nodes of displacements vs. time for three variants of loading

Failure indices analysis

Figure 8 show examples of contours of the second index in three laminas with different
reinforcement in A and B zones of BS, TS segments with the flanges. The second failure
index has the higher value in both zones of panel. This failure index is connected with
composite compression in the longitudinal direction of the fabric-reinforced laminae, tension
or compression in the transverse direction for mat or roving cut reinforced layers.

11
 Fig. 8. Failure indices of the individual layers of laminates (F - STR600, M - CSM450, R- CGR)

Figure 9 plotted the second failure indices values in selected laminas in zones A and B of
the top and bottom segments under full sequence of loading. The distributions of the failure
indexes have similar character, although values FI2 for top segment are slightly higher.

BS zone A layer 1 BS zone A layer 2 BS zone A layer 3

TS zone A layer 1 TS zone A layer 2 TS zone A layer 3

12
 BS zone B layer 2 BS zone B layer 3 BS zone B layer 4

TS zone B layer 2 TS zone B layer 3 TS zone B layer 4

Fig. 9. Selected contours of 2nd failure index (FI2) in subsequent laminas of both segments with flanges
(zone A: layer 1- CSM450, layer 3- CGR, layer 4: STR600; zone B: layer 2, 4- STR600, 3- CGR))

From figures 8-9 and other results it can be observed that:

1) According to failure indices distributions, applied maximal values of loading not cause
failures of composite covering segments. However, at FI2 contours in zone B of
segments, the cracking initiators in the fabric reinforced layers can be observed.
2) Maximum failure indexes in the fiber reinforced layers are equal 0.088-0.107 in zones
A and 0.074-0.09 in zones B. For the Hashin Fabric criterion, the strength ratio at the
integration points is equal n = 1 / 0.28 = 3.57 > nmin = 3 . This value satisfies global
standard equation [31].
3) The mat/roving layers have higher values of failure indices Maximum failure indices
in the roving reinforced layers are equal 0.188-0.192 in A zones and 0.134-0.163 in B
zones of the BS, TS segments. For the Maximum Strain criterion, the strength ratio is
equal 5.29. Moreover, maximum failure indexes in the mat reinforced layers are equal
0.171-0.173 of in the A zones of the segments. For the Maximum Strain criterion, the
strength ratio is equal 5.85.
4) Maximum average values of the failure indices are as follows: FI1 = 10,2% (CGR,
TS, Zone B), FI2 = 19,2% (CGR, TS, Zone A), FI3 = 2,8% (STR600, BS, Zone B),
FI4 = 5,7% (CGR, TS, Zone B), FI5 = 4,8% (CGR, TS, Zone B), FI6 = 9,0%
(STR600, BS, Zone B).
5) The dominant failure mechanism, observed in the covering system, corresponds to the
failure index FI2 (tension or compression in the perpendicular direction).

13
 6) Non-linear effects in the analyzed covering are minor due to a relatively low level of
loading.

4. Conclusions

The study presents FE numerical modeling and engineering calculations of rectangular

tanks composite covers, wherein:
a) MSC. Marc software was selected as computing environment due to available models
and options,
b) the geometry of the segments includes shells and structural laminates, i.e. glass –
polyester laminates of any sequence of layers,
c) the modeling includes joints of covers segments (rivet nut – bolt and anchor – bolt),
d) the calculations include fixed and variable loads and temporary standardized design
rules.
The study considers a basic type of composite rectangular tanks covers, i.e. composite
covers compound of a rectangular two-wave segments.
Presented FE numerical modeling and structural analysis of the composite covering are in
accordance with the following conditions:
a) geometric and material modeling covering and joints are compatible with the real
structures,
b) the number of degrees of freedom of the numerical model allows calculations to be
performed in a relatively short CPU time (less than one hour), owing to using the symmetry or
quasi-symmetry planes,
c) numerical models have been successfully validated experimentally,
d) numerical models should be possible to determine the design factors (safety),
e) methodology of covering composite structural analysis is compatible with current
standards loads,
f) modeling of joints of laminate segments with each other and with the covered object
includes elastic features of joints, clearances, contact and static friction.
FEA analysis of composite tank covers behavior shows the need for using proper design
rules to these structures. In the covers designing, it is required to take into consideration the
joints between the segments and the wall and the anisotropy of the composites. Tank covers
variety and applications possibilities result in the design of particular solutions. Full analysis
of the loading conditions and sequences of composite layers allow optimization and design of
safe structures. In the future works, attention will be paid to the covering system self-weight
minimize.

Acknowledgements
The study has been supported by the National Centre for Science, Poland, as a part of
research project No. N N506 1228 40, realized in 2011-2013. This support is gratefully
acknowledged.

14
Appendix. Selected failure criteria for composite materials

In the case of Hashin-Fabric criterion [28], at each integration point, failure indexes are
calculated according to formulas taking into account the failure of the matrix and fabric:
 tension of fibers in direction i, σ i > 0
   2    2    2 
FIj   i    i 2    i 3   , (A1)
 Ri t   S i 2   S i 3  
 
 compression of fibers in direction i, σ i < 0 :
   2    2    2 
FIj   i    i 2    i 3   , (A2)
 Ri c   S i 2   S i 3  
 
where i=1,2; j=1,2,3,4
 tension of matrix, σ 3 > 0 :
  2 2 2
   12    13    23 
2

FI 5   3           , (A3)
 R3t   12   13   23 
S S S 
 compression of matrix, σ 3 < 0 :
   2    2    2    2 
FI 6   3    12    13    23   , (A4)
 R3c   S12   S13   S 23  
where σ i ,σ ij - components of the stress tensor.
In the Maximum Strain criterion, MSC. Marc software [28] calculates the failure indices at
each integration point according to formulas:
  Si 
   for  i  0
 e
FIi    it  i  1,2,3;
  Si  for   0
 eic  i
(A5)
  ij 
FIij   , ij  12,23,31.
 g ij 
 
where: εi , γij - strain tensor components.
A laminate layer is described by 9 effective elasticity constants [7, 8]:
Ei – Young’s modules i, where i = 1,2,3
νij – Poisson’s ratios in principal planes ij, where i, j = 1,2,3 ,
Gij – shear modules in principal planes ij, where ij = 12,23,31 .
and 18 efficient strength constants:
Rit , Ric – tensile and compressive strengths in principal directions i, where i = 1,2,3 ,
S ij – shear strengths in principal planes ij, where ij = 12,23,31 .
eit , eic – ultimate normal strains at tension and compression in principal directions
ij  12,23,31
gij – ultimate shear strains in principal planes ij, where ij = 12,23,31 .

15
References

1. S. B. Sadineni, S. Madala, R.F. Boehm, Passive building energy savings: A review of

building envelope components, Renewable & Sustainable Energy Reviews -
Renewable and Sustainable Energy Reviews, 01/2011; 15(8), 3617-3631.
2. H.T. Wong, J.G. Teng, Buckling behaviour of model steel base shells of the Comshell
roof system, Journal of Constructional Steel Research, Volume 62, Issues 1–2,
January–February 2006, 4–19.
3. J.J. del Coz Díaz, P.J. García Nieto, J.A. Vilán Vilán, F.P. Alvarez Rabanal,
A. Navarro-Manso, M. Alonso-Martínez, Nonlinear analysis of the pressure field in
industrial buildings with curved metallic roofs due to the wind effect by FEM, Applied
Mathematics and Computation, Volume 221, 15 September 2013, 202–213.
4. R. Stuart, L. Shipley, A. Ghose, M. Hiremath, Thermal degradation of the concrete
roof of high-level waste storage tank, Computers & Structures - COMPUT STRUCT
01/1997; 64(5), 959-972.
5. L. A. Godoy, S. Lopez-Bobonis, On the Collapse of a Reinforced Concrete Digester
Tank, Thin-walled structures International conference; 3rd, Thin-walled structures;
edited by J. Zaras, K. Kowal-Michalska, J. Rhodes, 669-676.
6. G. Portela, L.A. Godoy, Wind pressures and buckling of cylindrical steel tanks with a
dome roof, Journal of Constructional Steel Research, Volume 61, Issue 6, June 2005,
808-824.
7. F. Trebuňa, F. Šimčák, J. Bocko, Failure analysis of storage tank, Engineering Failure
Analysis, Volume 16, Issue 1, January 2009, 26-38.
8. API Standard 650 - Welded Steel Tanks for Oil Storage, 10th Edition, 1998.
9. L. Zdravkov, T. Dincheva, Design of self-supporting dome roofs, 09/2011; In
Proceedings of: Eurosteel 2011, Budapest, Hungary.
10. Covering Systems – Technical Concept, C.F. Maier Europlast GmbH & Co KG,
Königsbronn, Germany, 2006.
11. Report 27- Covering Systems, C.F. Maier Europlast GmbH & Co KG, Königsbronn,
Germany, 2013.
12. A. Bondyra, Modeling and numerical study of composite covering of rectangular
tanks, Ph.D. Thesis, supervisor: M. Klasztorny, Military University of Technology,
Warszawa, 2012.
13. D. Nycz, A. Bondyra, M. Klasztorny, P. Gotowicki, Numerical modelling and
simulation of the composite segment bending test and experimental validation,
Composites Theory and Practice,12:2 (2012), 126-131.
14. de Beer J. and Walke, M., The design of GRP self-supporting covers for large circular
tanks and vessels., Composites Africa 2002.
15. Sz. Lutomirski, M. Lutomirska, The analysis of the origins of failure of the covers for
sewage tanks, XXV Scientific-Technical Conference Building Failures 2011,
Międzyzdroje 24-27 05 2011,: 591-600, [in Polish].
16. D.V. Scott, Advanced Materials for Water Handling: Composites and Thermoplastics,
Elsevier Science and Technology, 2000.
17. J. C. Thompson, C. S. Lee Parametric Studies of the Factors Affecting the Stability of
Segmental Dome Structures, [in] Trends in Structural Mechanics, Solid Mechanics
and Its Applications, Volume 54, 1997, 165-173.
18. Design workshop of the Oxygen Stability Chambers Waste Water Treatment Plantin
Goleniow, BP KE „ERPRO” Sp. z.o.o., Rybnik 2011, available by ROMA Sp. z.o.o.
Grabowiec).

16
19. I.M. Daniel, O. Ishai, Engineering mechanics of composite materials, New York –
Oxford, Oxford Univ. Press, 1994.
20. M. Klasztorny, A. Bondyra, R. Romanowski, D. Nycz, P. Gotowicki, Numerical
modelling and simulation of RN-B composite joint tensile test and experimental
validation, Composites Theory and Practice, 12:3 (2012), 198-204.
21. M.A. McCarthy, C.T. McCarthy, V.P. Lawlor, W.F. Stanley, Three-dimensional finite
element analysis of single-bolt, single-lap composite bolted joints: Part I – model
development and validation, Composite Structures 71 (2005), 140-158.
22. C.T. McCarthy, M.A. McCarthy, Three-dimensional finite element analysis of single-
bolt, single-lap composite bolted joints: Part II –effects of bolt-hole clearance,
Composite Structures 71 (2005), 159-175.
23. P.J. Gray, C.T. McCarthy, A global bolted joint model for finite element analysis of
load distributions in multi-bolt composite joints, Composites: Part B 41 (2010) 317-
325.
24. C.T. McCarthy, P.J. Gray, An analytical model for the prediction of load distribution
in highly torque multi-bolt composite joints, Composite Structures 93 (2011) 287-298.
25. S.B. Dong, F.K.W. Tso, On a laminated orthotropic shell theory including transverse
shear deformation, J. Applied Mechanics, Trans. ASME 39 (1972) 1091-1096.
26. E. Reissner, Small bending and stretching of sandwich-type shells, NACA Report No.
975 (1950) 483-508.
27. N.J. Pagano, Stress fields in composite laminates, Int. J. Solids & Structures 14 (1978)
385-400.
28. Marc 2008 r1, Vol. A, Theory and User Information, MSC.Software Co., Santa Ana,
CA, USA.
29. S.W. Tsai, Theory of composites design, 4th Edn., Think Composites, Dayton, 1987.
30. BS EN 13121-1:2003GRP tanks and vessels for use above ground. Raw materials.
Specification conditions and acceptance conditions.
31. BS EN 13121-2:2003 GRP tanks and vessels for use above ground. Composite
materials. Chemical resistance.
32. BS EN 13121-3:2008 GRP tanks and vessels for use above ground. Design and
workmanship.
33. BS 4994-1987 Specification for design and construction of vessels and tanks in
reinforced plastics.
34. BS 7159-1989 Code of practice for design and construction of glass-reinforced
plastics (GRP) piping systems for individual plants or sites.
35. BS 13923-2005 Filament-wound FRP pressure vessels - Materials, design,
manufacturing and testing.
36. R.M. Jones, Mechanics of composite materials, London, Taylor & Francis, 1999.
37. ASTM D 5379/D 5379M-98 Standard Test Method for Shear Properties of Composite
Materials by the V-Notched Beam Method.
38. EN ISO 527-4:1997 Plastics - Determination of tensile properties - Part 4: Test
conditions for isotropic and orthotropic fibre-reinforced plastic composites.
39. Marc 2008 r1, Vol. B, Element Library, MSC. Software Co., Santa Ana, CA, USA.

17