See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/353193655

ScienceDirect Structural Design of the Base of an Internal Combustion Engine

Using FEM

Preprint · March 2021

CITATIONS READS

0 341

2 authors:

Carlos Cuenca Cabrera Luis Gonzaga Bermeo

Escuela Superior Politecnica del Litoral Escuela Superior Politecnica del Litoral
11 PUBLICATIONS 24 CITATIONS 2 PUBLICATIONS 2 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Carlos Cuenca Cabrera on 12 July 2021.

The user has requested enhancement of the downloaded file.

 Available online at www.sciencedirect.com

ScienceDirect

[Name of the proceedings]

Structural Design of the Base of an Internal Combustion Engine

Using FEM
Carlos A. Cuencaa,b,1, Luis Gonzaga B.c
a
Escuela Superior Politecnica del Litoral, Facultad de Ingenieria Mecanica y Ciencias de la Produccion, Guayaquil 09015863, Ecuador
b
Universidad Regional Amazonica Ikiam, Facultad de Ciencias Socio Ambientales, Tena 150150, Ecuador
c
DINAVALE, Daule 090602, Ecuador

Abstract

This document aims to show a rational process to estimate the load generated by the vibratory torque of an Internal Combustion
Engine (ICE). As consequence, being able to design the structure of the engine bed. In this way, the objectives are to determine the
equivalent load produced by the components of the ICE at a certain working speed while the use of Finite Element Method software
let to determine the minimum structural dimensions required to stand these loads. In order to determine the cyclical load to be
applied, the mechanical characteristics of the engine were used. By other hand, ANSYS - Workbench was used for the structural
optimization in which a geometrical model of the structure, generated in Rhinoceros, was proved at different structural
combinations of thickness until obtaining one that satisfies the requirements. Therefore, the scantling of the primary and secondary
stiffeners that engine base have to withstand due to cyclical torque load are obtained.
[copyright information to be updated in production process]

Keywords: Computer aided engineering, Structural design, Engine torque, Engine base, Finite element method

1. Introduction

Nowadays there are different ways to design structures in a rational form. The optimization of variables such as
weight and load capacity allow to achieve the design requirements of the structure. One of the accepted methods is the
use of the ANSYS platform in the Workbench environment (ANSYS WB), which is based on finite element analysis
[1].
The use of CAE technology allows to obtain stresses, product of static or dynamic loads that act on a structure, as
well as the deformations, allowing to the design engineer to select the criterion of safety factor that is convenient for
the design of the structure. Being this an accepted and suggested methodology in the Industry of Structural Design,
[2], [3].
To carry out the structural analysis using this methodology, a process of five stages are stated: definition of the
problem, elaboration of the geometric model, elaboration and analysis of the finite element model, analysis of
results, and experimental validation [4], [5], the first four (4) stages are carried out using a computer and the last one
using the real prototype to validate the information obtained from the numerical model. This process can be applied to
perform the analysis and design of simple and complex structures as shown in [6] and [7].
Therefore, in order to continue this study, it is necessary to know, among other data, the main dimensions and
structural components of the studied area. In addition, it is important to identify which external phenomena produce
relevant loads. As consequence, it will allow to build the geometrical model as well as the numerical model easily. In

1 Corresponding author.
E-mail address: [email protected]
this way, it is defined that the analysis will be carried out in the machinery room of a fishing vessel, specifically the
structure on which the main engine is supported, called the engine foundation.
This structure is modeled using surface-type geometric elements from CAD software. Then it is exported to ANSYS
WB, where the material is implemented, and the number of elements into which the structure is divided is settled.
After that, the mesh generated is tested at different size of elements in order to demonstrate independence of mesh
sizing. To complete the numerical model of the structure, the dynamic and static induced loads on the engine
foundation are applied. These are produced mainly by the propulsion system of the ship, principally by the propeller,
propulsion shaft and gearbox – motor, static pressure, slamming pressure, bending moment due to wave and still water.
Considering the dynamic loads, only the cyclical torque generated by the engine is calculated, since the “hit”
pressure will rarely affect the hull. Likewise, for the static loads, the hydrostatic pressure which acts on the modeled
hull section was allocated, as well as weight of the engine on the withstanding structure. For the last loads, the bending
moment in swell and calm water conditions are not considered since only a small section of the structure was modeled.
Finally, using the characteristics of the vessel and the propulsion system, the loads are applied to the mesh where
the engine is in contact with the structure. In addition, the thickness of structure that make up the foundation of engine,
is evaluated at different combination of scantling, specifically for webs and flanges to obtain the minimum thickness
at which the structure stands the loads.

2. Materials and method

2.1. Structure Structural Description and Main Dimensions of the Engine and the Ship

Considering that the present study uses ANSYS WB to analyze the structure that works as foundation for a marine
engine, Cummins QS8.3-500, installed in an ecuadorian fishing vessel of 27 meters length overall and floating at
design draught of 2.7 meters. The main particulars and data such as Power vs Rotational speed are shown in Table 1
and Table 2, [8].

Table 1. QSC8.3 Cummins engine main features.

Engine Specifications
Configuration In-line 6-cylinder, 4-stroke diesel
Bore & Stroke 114 mm x 135 mm
Displacement 8.3 L
Aspiration Turbocharged / Aftercooled
Rotation Counterclockwise facing flywheel
Max Torque 1799 N-m
Weight 896 Kg

Table 2. QSC8.3 Cummins engine power / speed ratings.

RPM BHP
2000 474
2200 490
2400 493
2600 493

In addition, the area of study in the machinery room has a length of 4.50 m and has a structural distribution as
shown in Fig. 1, in which it is observed that the base of the engine is made up of two (2) longitudinal stiffeners (profile
built) type "L" (84 x 26 cm), the thickness web and flange are values to be determined.
 Fig. 1. 3D view of the structural distribution of the engine area in the analyzed ship.

Other structural arrangement is described in Table 3.

Table 3. Structural arrangement at the machinery room of fishing vessel.

Web Flange Thickness

Item Element
(mm) (mm) (mm)
Keel Plate 300 --- 14
Longitudinal stiffeners Plate 100 --- 8
Main frames L-shape 540 70 6
Intermediate frames L-shape 460 70 6
Transversal intermediate stiffeners L-shape 370 100 6
Shell plate Plate --- --- 12

The scantling of the preliminary engine base is fundamental in the elaboration of the geometric model of the
structure as well as the numerical model.

2.2. Geometrical Modeling Generation for the Structure

Using this information, it is proposed to follow the previously defined scheme to perform an analysis with CAE
tools. Hereby, the geometric model is built using a CAD tool that allows to generate the hull shape and structural
arrangement to be analysed. Rhinoceros is the software chosen for the present study, since it is possible to model the
structure using surfaces. After that, exported to ANSYS WB, [9], and modified in SpaceClaim (SC) module, [1].
Within this module, corrections are made in the geometric model to avoid later faults in the mesh of the structure.
Only half of the study area has been modelled, taking advantage of the symmetry with respect to the plane where the
keel of the vessel is located at the center line. Once the model has been corrected in the SC module, it must verify that
the shared edges between several surfaces are properly connected, which is denoted by a blue colour for shared edges
and red for free edges, as shown in Fig. 2.
 Fig. 2. Geometric model corrected in SpaceClaim module.

Once imported the numerical model into the ANSYS, the material is added, A131 Steel (naval steel) [10]. Besides,
the mesh is generated with an adequate number of elements to later apply load conditions and perform structural
analysis. This allows to obtain the response of system. As consequence, this permits the selection of the thickness for
the structural elements that make up the structure without failure.
Once the mesh is generated, Fig. 3, it shows a uniform mesh with a total of 19168 and 19116 nodes and linear
elements, respectively. Then, the analysis of quality for the element of the mesh shows that 99% of elements have an
aspect ratio value of 1.29 which allows to continue with the analysis, due to the value that denotes a better quality of
mesh is 1, [11].

Fig. 3. Model mesh generated in ANSYS Workbench.

Likewise, the thickness values are added to structural elements of the geometry modelling, see Table 3. It is possible
to visualize, according to the thickness defined for them, as shown in Fig. 4.
 Fig. 4. Thickness definition, in mm, on the geometric model classified by colors on Workbench.

2.3. Numerical Modeling Generation for the Structure

Finally, to define the numerical model in which the stress analysis and subsequent thickness selection will be carried
out, the load generated must be determined and how it will be applied on the modelled engine bed. For the analysis of
the applied load, it is known that the propulsion system of a boat may generates loads of forces and moments through
three (3) main elements such as: the engine supported on the structural base, the axis that transmits the thrust, and the
non-uniform thrust generated by the propeller as shown in Fig. 5, [2], [12].

(a)

(b)
Fig. 5. Propeller, Shaft and Engine induced actions, a) 2D view; b) 3D view.

In the present study, only the load generated directly by the torque of the engine acting on the structure that supports
it has been selected as a dynamic load to resize its structural elements if required. Furthermore, for a more exhaustive
study it is suggested to also include the loads generated by the propeller and the shaft of the propulsion system. In
addition, the load due to the engine mass and the hydrostatic pressure can be directly added to the geometric model
due to the linear nature of the analysis.
 Then, to determine rationally the load generated by the torque of engine, the process shown in Fig. 6 is performed
[13]. This process indicates how to determine the load generated by the torque of the engine, once its dimensions and
geometry characteristics are known, whether the engine is 4 or 2-strokes. See information shown in Table 1and Table
2. With these values the Power and the pressure indicated at the nominal working speed of the engine with equations
1 and 2 are determined.

Fig. 6. Rational process to determine the load due to engine torque.

𝑃𝑒𝑓𝑓 (1)
𝑃𝑖 =
𝜂𝑀
 𝑃𝑖 (2)
𝑝𝑖 =
𝐴·𝑆·𝑍·𝑁·𝐾

Where 𝑃𝑒𝑓𝑓 is the effective power of engine, 𝜂𝑀 is the efficiency, 𝑃𝑖 is the indicated power, 𝐴 is the cross-section
area of cylinder, 𝑆 is the stroke, 𝑍 is the numer of cylinder in engine, 𝑁 is a working rate speed in RPM and 𝐾 is a
convertion factor which for a 4T engine is 0.5. Then, using equation 3 obtained from [14] it is possible to determine
the tangential forces as a function of the indicated pressure of a 6 – stroke internal combustion engine.

𝜏 = 0.11𝑝𝑖 𝑇𝑒𝑜 + 1.21 (3)

Finally, to obtain the value of the equivalent force applied, the value of Torque generated is related to the distance
between engine supports using equation 5, that is, the separation between the two girders of foundation of engine of
the analyzed vessel.

𝑇 (4)
𝐹𝐴𝑝𝑝 =
𝑑

Where 𝑑 is the distance between main frames and 𝐹𝐴𝑝𝑝 is the equivalent applied force. This allows to obtain an
equivalent torque of 123.8 Kg·m when the engine works at 2000 RPM and considering structural arrangement of the
engine foundation which have a span of 1.21 m, then the equivalent load corresponds to 102.31 Kg which is distributed
between the points of application of the analyzed structure along the engine supports of the structural base.
In this way and considering that the engine has a length of 2.2 m., the load has been distributed in two (2) zones,
called Nodal Force 1 and Nodal Force 7, so 51.2 Kg (511.7 N) are applied at each loading zone. These loads are
applied at frequency of 100 CPS which is equivalent to working rate speed, 2000 RPM.
By other hand, the hydrostatic pressure which acts on hull surface is considered acting from keel to design water
line at 2.7 m. Therefore, the pressure is possible to distribute along the hull as shown in Fig. 7. The density of water
used for calculation is 1025 𝑘𝑔/𝑚3 .

Fig. 7. Distribution of applied hydrostatic pressure in the Numerical Model in ANSYS Workbench.

The load due to the weight of the engine is distributed among its four (4) supports, summing a total load of 4390
N (448 kg) at each support over the engine foundation, Fig. 8.
 Fig. 8. Applied loads and boundary conditions in the Numerical Model in ANSYS Workbench.

For the boundary conditions of the numerical model, a simply supported system has been considered, due to in the
limits of the geometric model there are other structural elements that prevent the existence of exaggerated
deformations. Furthermore, the applied load will directly affect the surrounding structural area of engine.

3. Results

Once the loads and boundary conditions have been defined, the corresponding iterations are carried out to obtain
the value of the minimum thicknesses to withstand the load generated by the engine torque when it is working at
nominal speed. For this, the thickness variable of the main structural element that forms the base of the engine is
located as a parameter "P", which allows the initial and final thickness values to be located around where we want to
obtain the stresses. Finally, using Von - Misses failure criteria, the safety factor is sought for the maximum stresses of
the structure not less than 1.5, [4], [15].
Then, the value of stress which is entered in S-N curve of the structural detail, permits to know the life cycle. This
value must be contrasted with design life of the structure, which must be 25 years. This value can be determined by
performing fatigue analysis of the structural detail.
When performing one of the tests with the applied load and considering that the thickness is 5 mm for the foundation
of engine, the response forces for the flange and the web are shown in Fig. 9. There, it can be seen that the area of the
supports that reflects the highest response, corresponds to the load applied in Nodal Force 7 that is located between
two transverse frame of hull, while Nodal Force 1, due to it is applied directly on a structural zone formed by
longitudinal and transverse frames, does not meet a high response.
Once found that the model can generate results correctly, the process of parameterization of the variable already
described in previous paragraphs is implemented. In this case, the value of the thicknesses of the structure of the
engine bed is parameterized, being able to vary the values of the flange and the web either with the same value or with
combinations of thickness to identify which of them produces the least von-Mises stress.
 Fig. 9. Stress response duo to applied load on the engine foundation.

Thus, Table 4 shows different values of thickness with response stresses as well as the Safety Factor (SF) and
considering 235 MPa as a yield stress, [10].

Table 4. Von-Mises stress response for different thickness assumption.

tweb tflange Stress VM Safety

(mm) (mm) (MPa) Factor
1 1 405 0.6
2 2 132 1.8
3 3 79 3.0
4 4 59 4.0
5 5 48 4.9
6 6 42 5.6
7 7 38 6.2
8 8 35 6.8
9 9 32 7.3
10 6 40 5.8
10 10 30 7.8
12 6 40 5.9
12 8 33 7.1
12 12 27 8.8

In Fig. 10 is possible to view stresses at each combination of thickness for web and flange. From the values
obtained, it can be observed that the effect of the load generated by the torque of engine, when it is applied, does not
produce responses to worry about. Besides, for the static applied load, the thickness of 2 mm for both, the flange and
the web that forms the base of the engine already allows a safety factor of 1.8.
 Fig. 10. Stress response due to applied load on engine foundation for each combination of 𝑡𝑤𝑒𝑏 and 𝑡𝑓𝑙𝑎𝑛𝑔𝑒 combination.

Then, to include the dynamic effect of the load, a fatigue analysis is performed with the response stresses obtained
on the structural detail of interest. For this calculation it is required to know the Stress – Cycle of life (SN) curve of
the material and the detail to be analyzed subjected to a certain type of load, [16]. In Fig. 11, the two stresses response
are input in the curve and straightforward is possible to know the life of the two combined thickness of specimen.

Fig. 11. Curve S-N for the combination of thickness in web and flange (12;6) and (12;12).

Table 5 shows the number of cycles that the structural detail supports for each combination of flange and web
thickness.
 Table 5. Cycles for different thickness combination.

( tweb ; tflange ) N
(mm; mm) (cycles)
(1;1) 2.67E+04
(2;2) 2.33E+06
(3;3) 2.48E+07
(4;4) 1.52E+08
(5;5) 4.80E+08
(6;6) 1.11E+09
(7;7) 2.06E+09
(8;8) 3.47E+09
(9;9) 5.49E+09
(10;6) 1.43E+09
(10;10) 8.30E+09
(12;6) 1.60E+09
(12;8) 4.49E+09
(12;12) 1.71E+10

Then, it is estimated that the vessel operates 28.6% of the time at 2000 RPM. Besides, it is possible to calculate the
value of the working cycles per unit of time at which it is under that regime, 𝑛𝑖 .

1
𝑛𝑖 = % 𝑟𝑒𝑔𝑖𝑚𝑒𝑛 · 𝑅𝑃𝑀 · 𝐾 · ( ) [𝐶𝑃𝑆] (5)
60

Finally, Miner's rule is used to determine the value of the damage 𝐷 produced by the load in each combination of
thicknesses, to then determine the time in which some damage occurs at the speed of the analysed working frequency.

𝑛𝑖
𝐷= [1/𝑠] (6)
𝑁

1
𝐿𝑇 = · (3.156𝐸 + 07) [𝑦𝑒𝑎𝑟𝑠] (7)
𝐷

Where the 𝐿𝑇 is the Lifetime what means the useful life time of a structural detail. The lifetime values for each of
the analysed combinations are shown in Table 6¡Error! No se encuentra el origen de la referencia.. From Table
6¡Error! No se encuentra el origen de la referencia. it is observed that only the combination of 12 mm thicknesses
for both elements, flange and web, permit obtaining a structural useful life value greater than the 25-year design useful
life suggested in [16]. Likewise, it can be identified that for the combination (12; 6), the structure only has a useful
life of three (3) years, even though in the previous static analysis it has a security factor, 𝑆𝐹 > 1.5.
From here, it can be determined what is the importance of dynamic analysis for the design of this type of structures,
so it should be suggested to do the analysis using the load values but applied dynamically on the structure. These loads
should be applied to a corresponding working frequency, which will vary according to the RPM of the motor's working
condition in each situation.
 Table 6. Von-Mises stress response for different thickness assumption.

( tweb ; tflange ) D LT
(mm ; mm) (1/s) (years)
(1;1) 7.12E-04 0
(2;2) 8.17E-06 0
(3;3) 7.67E-07 0
(4;4) 1.26E-07 0
(5;5) 3.97E-08 1
(6;6) 1.71E-08 2
(7;7) 9.26E-09 3
(8;8) 5.49E-09 6
(9;9) 3.47E-09 9
(10;6) 1.33E-08 2
(10;10) 2.30E-09 14
(12;6) 1.19E-08 3
(12;8) 4.24E-09 7
(12;12) 1.11E-09 29

4. Discussion

To determine if the model is independent of the mesh size (mesh size effect), several models with different mesh
refinement were made and the two values of thickness combinations were chosen, (12;12) and (12;6), to validate if
there is any difference due to the refinement. In Fig. 12 it is observed that after 19000 nodes, the difference of response
keeps stable, obtaining a difference of 2% for the results obtained above that number of nodes.

Fig. 12. Mesh independency of the model.

Therefore, it can be determined that the number of elements and nodes chosen for the study is adequate.
The importance of dynamic analysis for structural design is observed. Likewise, it can be observed that there are
combinations of thicknesses for the structural arrangement in which the life cycle is shorter since they support a higher
load response.
To apply the load due to the bending moment for calm waters and in swell conditions, it is proposed to carry out
the numerical model of the whole machinery room in order to simulate the total effect that this type of load has on the
structure. In the present study, since only a portion of the engine room is modelled, not all the loads should be added,
but a percentage of it.
After using Miner's rule, the useful lifetimes for the analysed structural detail are known. This analysis makes it
possible to estimate the remaining life of the structural detail if due to some external effect the thickness of the material
decreases and coincides with the thickness combinations presented.

5. Conclusions

One section of the machinery room of fishing vessel, built in A131 naval steel, was analyzed to determine the
minimum dimensions of the structure due to the load generated mainly by the engine that is part of the propulsion
system. After modelling a section of the machinery room in Rhinoceros and export to ANSYS, where the geometry
was defined for analysis in the Static Structural module, the mesh was made up of 19168 and 19116 nodes and linear
elements, respectively.
To define the numerical model, the boundary conditions were assumed for this study as simply supported at the
elements that surround the structure. Then, the load due to torque was rationally calculated using the geometric and
operating characteristics of the installed engine Cummins QS8.3-500 that has a power of 500 BHP at 2000 RPM. Once
following the methodology described in Fig. 6. it is possible to obtain the equivalent torque produced by the engine
which is 123.8 Kg·m. Considering the geometry of the engine foundation and that it has 2 rear supports and 2 front
supports, then a force of 51.16 Kg (511.7 N) is applied at each support at 100 CPS. These loads were applied only
statically in the area equivalent to supporting the engine over the flange of base of engine.
Then, once the numerical model was completed, the variables of interest were parameterized, which are the web
and flange thicknesses of the engine base where in conjunction with the defined load in the numerical model, it
produces a corresponding stress response for each entered thickness combination, with values from 1 to 12 mm being
chosen.
It is observed that the minimum thickness for both the flange and the web of this structural element must be 12 mm
to get a higher value than the recommended design life of 25 years. Some other combination of thickness was analyzed
but they do not meet the requirement of design life.
To have a more exhaustive study, it is suggested to carry out a dynamic analysis of the response due to stress
generated by the slamming, that is, to apply an equivalent pressure at a working frequency. Also, bending moment
due to still water and wave conditions should be applied to get more accurate values of stress response. This will allow
to carry out a fatigue analysis that will help to define the desired thickness of the structure in a dynamic environment.

Acknowledgements

The authors of this study thank the developers of ANSYS for allowing students to conduct research in the student
module of the software.

References

[1] J. Swanson, “Ansys, 2020 R1.” NASDAQ, Canonsburg, Pennsylvania, 2020.

[2] Society of Naval Architects and Marine Engineers, “Ship Design and Construction,” in The Elements Of
Shipping, T. Lamb, Ed. Boston, MA: SNAME, 2003.
[3] Türk Loydu, “Hull structures and Ship Equipment,” in Rules for the classification of naval ships, no.
October, Estambul: Chamber of Marine Engineers and the Association of Insurance and Reinsurance
Companies of Turkey, 2013.
[4] K. Cheng, “Finite element analysis and structural optimization of the box on the ANSYS workbench,” Adv.
Mater. Res., vol. 211–212, pp. 434–439, 2011.
[5] iESSS, “Notas de aula: Análise numérica estrutural utilizando o método dos elementos finitos,” in Pós-
Graduação em Simulação em Simulação Computacional, Sao Paulo: Instituto ESSS de Pesquisa e
Desenvolvimento, 2018.
[6] M. Yungán and J. Marín, “Proceso De Modelización De Estructuras Complejas Para Análisis Con
Elementos Finitos,” Jornadas Técnicas sobre Diseño y Prod., pp. 1–7, 2010.
 [7] F. Macas Camacho and J. Marín López, “Análisis estructural de la embarcación fluvial Samarina utilizando
el Método de Elementos Finito,” vol. 1, no. 1, Guayaquil, 03-Mar-2009.
[8] Cummins Engines, “QSC8.3 | Cummins Inc.,” 2020. [Online]. Available:
https://www.cummins.com/engines/qsc83. [Accessed: 03-Dec-2020].
[9] R. McNeel and others, “Rhinoceros 3D, Version 6.0.” Robert McNeel & Associates, Seattle, WA., 2010.
[10] ASTM International, ASTM A131 / A131M - 19 Standard Specification for Structural Steel for Ships. West
Conshohocken, PA, 2019.
[11] A. Gay Neto, Método dos Elementos Finitos e Geração de Malhas. Sao Paulo: Instituto ESSS de Pesquisa e
Desenvolvimento, 2017.
[12] SNAME, Ship Design and Construction. New York: American Society of Civil Engineers, 1980.
[13] C. Carlos A., “Determinación de la vida útil del fondo de una embarcación construida en aluminio, por carga
vibratoria del sistema propulsivo,” Escuela Superior Politécnica del Litoral, 2014.
[14] V. Moreno, Motores de combustión interna. Madrid: Escuela Técnica Superior de Ingenieros Navales, 1980.
[15] P. Martí Montrull, “Aplicación de las técnicas de optimización en el diseño de estructuras ligeras para
grandes luces,” II Encuentro Internacional de Estructuras Ligeras para Grandes Luces: Second
International Meeting on Lightweight, Murcia, Apr-1995.
[16] DNV GL, “DNVGL-CG-0129: Fatigue Assessment of Ship Structures,” no. October, pp. 1–129, 2015.

View publication stats