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

net/publication/271972719

Model for dynamic behavior of the crankshaft of an air cooled diesel engine
subjected to severe functioning

Article  in  International Journal of Automotive Technology · August 2014

DOI: 10.1007/s12239-014-0086-7

CITATIONS READS

2 957

3 authors, including:

Oumar Keita Jamel Bessrour

École des Mines d'Albi-Carmaux University of Tunis El Manar - Ecole Nationale d'ingénieurs de Tunis
11 PUBLICATIONS   77 CITATIONS    51 PUBLICATIONS   436 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

friction stir welded View project

A Time Domain Boundary Integral Method for acoustic problems in an arbitrary uniform mean flow View project

All content following this page was uploaded by Jamel Bessrour on 23 February 2015.

The user has requested enhancement of the downloaded file.

International Journal of Automotive Technology, Vol. 15, No. 5, pp. 823−833 (2014) Copyright © 2014 KSAE/ 079−15
DOI 10.1007/s12239−014−0086−7 pISSN 1229−9138/ eISSN 1976−3832

MODEL FOR DYNAMIC BEHAVIOR OF THE CRANKSHAFT OF AN

AIR COOLED DIESEL ENGINE SUBJECTED TO SEVERE FUNCTIONING

O. KEITA1)*, T. HEDHLI2) and J. BESSROUR2)

1)
Building Architecture and Town Planning Department (BATir), Université Libre de Bruxelles (ULB), Bruxelles 1050,
Belgique
2)
Laboratoire de Mécanique Appliquée et Ingenierie, Ecole Nationale d'Ingénieurs de Tunis, BP 37
Le Bélvédère, Tunis 1002, Tunisie

(Received 15 February 2013; Revised 8 May 2013; Accepted 30 September 2013)

ABSTRACT−The dynamic behavior of the engine organs in severe conditions is complicated to identify. In this paper, the
dynamic behavior of the crankshaft of the diesel engine Deutz F8L413 direct injection-type air cooled in the severe operating
conditions is investigated in a 3D global model. The maximum operating characteristics of the engine are experimentally
measured on a bench test equipped with a hydraulic brake. The most stressed areas of the crankshaft are determined by
numerical simulations. In addition, an analysis of the fatigue behavior of the crankshaft is carried out by using two fatigues
criteria. The efficiency of the model is demonstrated by comparing between the numerical results and the experimental data
obtained with the natural modes of vibration test.

KEY WORDS : Crankshaft, Dynamic behavior, Fatigue, Vibrations, Severe functioning

1. INTRODUCTION versus number of cycles curves and to evaluate the fatigue

endurance limits of ductile iron crankshaft under crankpin
The automotive crankshafts are subject to a large number rolled and un-rolled conditions. One of the advantages of
of cyclic loading during their service, causing many engines with cylinders in V-type configuration is that they
fatigues and vibration problems. For the particular case of greatly reduce the overall engine length, height and weight
engines with cylinders in V-type, these dynamic compared to an equivalent inline configuration. It was
phenomena are more important and merit to be clearly found that if the same 6 cylinder inline engine is converted
identified in the context of maximum engine load, i.e.to into V-engine, the level of amplitudes of primary and
regimes corresponding to maximum power and torque. secondary moments is reduced up to 50% (Ide et al., 1990).
Many researches were carried out in the field of dynamic A system model analyzing dynamic behavior of IC engine
crankshafts based on simplified models because of the crankshaft has been developed in (Niinoa et al., 2002). The
complexity of the geometry and the many interactions model couples the dynamic structure of the crankshaft, the
involved. Large static and dynamic analyses were hydrodynamics lubrication of the main bearings and the
performed on the crankshaft to improve their durability rigidity of the crankcase. Recently, an experimental study
(Henry et al., 1992) and performance (Dimbale et al., on engine dynamics Model based in-cylinder pressure
2011). An algorithm is developed in (Nikolic et al., 2012) estimation was carried out in (Feilong et al., 2012) and (Yu
to obtain a theoretical wear diagram of IC engine et al., 2013). A more complicated multi degree-of-freedom
crankshaft main bearings by considering the crankshaft as engine dynamic model was used in the works of (Taraza,
beam. A simulation of crankshaft dynamic behavior of 2002; Schagerberg and McKelcey, 2003; Larsson and
engine with cylinders in V type was performed in (Niinoa Schagerberg, 2004) to estimate in-cylinder pressure by
et al., 2002) by measuring V-type experimentally using utilizing a crankshaft integrated torque sensor. However,
gauge the amplitude of stress at the crankpins for high and only few researches has been conducted on 3D models of
low engine speed. The research in (Çevik and Gürbüz, the crankshaft, especially on engines with cylinders in V-
2012) proved the effect of crankpin rolling on the fatigue type under maximum loading conditions (regime of
behavior of ductile cast iron crankshaft used in diesel maximum power and torque). These real conditions of
engines. Based on the staircase test methodology, they maximum loading allow realistic prediction of dynamic
performed resonant bending fatigue tests to obtain stress stress levels which is very important for durability,
lightness and optimization of the dynamic responses (Kim
*Corresponding author. e-mail: okeita@ulb.ac.be and Woo, 2013). It is also important for resistance to

823
824 O. KEITA, T. HEDHLI and J. BESSROUR

fatigue as shown in (AR et al., 1989). conducted in laboratory to identify the eigen modes of free
This paper investigates a 3D model using the maximum crankshaft (without flywheel) and compare to the eigen
loading conditions deduced from experimental measure- modes from numerical simulations.
ments on bench test to identify the dynamic behavior of a
diesel engine crankshaft with cylinders in V-type taking as 2. EXPERIMENTAL STUDY ON BENCH TEST
a case study Deutz diesel engine types F8L413 with four
(4) crankpins to 90 widely spread. This engine, air-cooled, The aim of bench test is to measure the operating
is heavily loaded. The novelty of the present model is that parameters of the engine, in order to: (i)-plot external
the parameters are derived from engine real working characteristics curves (power, torque, specific consumption)
conditions which are used to identify the dynamic behavior according to crankshaft number of turns; (ii) determine,
of the crankshaft. This paper is organized as follows. First, through the connecting rod the acting efforts on the
identification of optimal regimes (maximum power and crankpins. The used fundamental material is a bench test
torque and minimum specific consumption) is made with equipped with a hydraulic brake. The test conditions and
experimental data collected on bench test. It is followed by other materials prepared are described in (Taher, 2003).
the kinematic and dynamic characterization of the piston Figure 1 (a, b and c) show the curves of power, torque and
and the connecting rod this, is to determine the efforts specific consumption, respectively.
exerted on the crankshaft through the connecting rod-crank From these performance curves, it is find that 2650 tr/mm
handle system at maximum identified regimes. Then corresponds to the maximum power engine; 1600 tr/mm is
hydrodynamic bearing-crankshaft contact model is minimum consumption and 1400 tr/mm is maximum
presented as well as the 3D finite element modeling of the torque. The crankshaft dynamic behavior is studied on those
crankshaft. The simulations results and post-processing are different regimes numerically by the finite element method.
presented. The fatigue calculations are performed. The
vibration modes were also identified. Vibration tests were 3. KINEMATICS AND DYNAMICS
CARACTERIZATION

3.1. Kinematic Characterization

The objective of this section is to determine the linear and
angular accelerations of the connecting rod and the piston
which is used in the crankshaft dynamic characterization.
Figure 2 shows the technological diagram and parameteriza-
tions of the connecting rod-crank handle system.
The reference (O; X; Y; Z) is defined by the axis (X; Y)
of the cylinder and the axis (O; Z) of the crankpin crank
handle (Figure 2). It is assumed that the pulsation ω is
practically constant (case of multi-cylinder engines). We
denote by r the crank handle radius (OA), l is the length of
connecting rod (AB) and λ = r/l=0.2632. Displacement of
piston is y = OB. Projecting the contour of OAB on the axis
(O; X), we obtain the relation between

θ and ϕ as follows:

sinϕ = −λ sin θ (1)

Figure 1. Engine regime: (a) power; (b) torque; (c) specific Figure 2. Technological scheme and parameterizations of
consumption. connecting rod-crank handle system.
 MODEL FOR DYNAMIC BEHAVIOR OF THE CRANKSHAFT OF AN AIR COOLED DIESEL ENGINE 825

taking into account that cosϕ ≥ 0(see Figure 2), we get: momentum that is a function of the moving parts and their
accelerations. To calculate this momentum, it is necessary
cos ϕ = 1 – λ 2 sin2 θ (2) to determine the mechanical actions applied to the
connecting rod head and feet. The engine types F8L413
3.1.1. Kinematics of the connecting rod Deutz diesel is an engine with 8-cylinders in V-type. At
By deriving equation (1) and taking into account equation each of the four crankpins are mounted to two head of
(2) we obtain angular acceleration of the connecting rod as connecting rod (figure 3). A multi-cylinder engine of V-
2 type is constituted by two inline engines offset by an angle
λω sinθ -
ϕ·· = ------------------------------------
3⁄2
( 1 – λ2 ) (3) δ = (Y1,Y2) called the V opening angle (figure 3) (Swoboda
( 1 – λ sin 2 θ )
2
et al.,1998). In our case V opening angle δ = 4π/8 = π/2.
The linear acceleration of the connecting rod is therefore Let FA1 and FA2 be the forces exerted by the crankpin on the
expressed by following relation: connecting rods (b1) and (b2) at point A, respectively and
FB1, and FB2 the forces exerted by the pistons on the
Γ ( G ) = ( ω 2 r sinθ – lG ϕ·· cosϕ + l G ϕ· sinϕ )x –
2

(4) connecting rods (b1) and (b2) at points B1 et B2

( ω 2 r cosθ + l ϕ·· sinϕ + l ϕ· cosϕ )y
2
G G respectively.
The detailed calculations of relations (3) and (4) are
explained in Appendix A. 3.2.1. Case of the connecting rod (b1)
The reduction elements of the torsor quantities of
3.1.2. Kinematics of the piston accelerations of the connecting rod (b1) at its gravity center
Projecting the contour of OAB on the (o, y) and taking into G1(which is also its inertia center) are δG1(b1) and mb1Γ(G1)
account equation (1), we obtain the acceleration of the mb1, is the mass of the connecting rod (b1), XA1 YA1, XB1 and
piston as follows: YB1 are projections of FA1 and FB1 respectively on (O;X1) and
2 (O;Y1) axes (Figure 4). By applying the fundamental
λ sin θ
y·· = –ω r cos θ – ------------------------------------ + λ ( 1 – λ sin θ )
2 2 2 3⁄2
(5) principle of dynamics to the connecting rod (b1), and
( 1 – λ sin2 θ )
2
projecting the equations on the different axes, we obtain the
following equations:
3.2. Dynamic Characterization
Reduction elements of the mechanical actions applied to mb1 .Γ ( G1 ).X1 = m b1 .x··G1 = XA1 + XB1 (6)
crankshaft crankpin and the piston are reduced on the axis
of the crankpin and the connecting rod to a force and a mb1 .Γ ( G1 ).Y1 = m b1 .y··G1 = YA1 + YB1 (7)
lG1 cosϕ 1 .m b1 x··G1 + lG1 sinϕ1 .mb1 .y··G1 −
(8)
lY sinϕ – l.X cosϕ = I ϕ··
B1 1 B1 1 G1Z1 1

3.2.2. Case of the piston of the connecting rod (b1)

By isolating the piston from the system, as in Figure 4, and
applying the fundamental principle of dynamics we obtain:

FB1 + P + N + Ff = mp.Γ (B) (9)

where FB1, is the action of connecting rod on the piston. N,

Figure 3. Defining reference for a engine with V- type

cylinder: (a) Technological scheme; (b) Calculation scheme. Figure 4. Calculation scheme for connecting rod (b1).
826 O. KEITA, T. HEDHLI and J. BESSROUR

is the action of cylinder on the piston (reaction of the

jacket). P, is the resulting combustion gas pressure uniformly
distributed over the head of the piston and expressed as
follows:

– PY 1 = –P ( θ ).S.Y1 (10)

in which P(θ) represents differential pressure on both sides

of piston. Ff, is the frictional force between the jacket and
the piston rings and is expressed as:
·
Y1- N ( M )
Ff = – f ------- (11)
·· Figure 5. Force exerted by connecting rod (b1) on the
Y1 crankpin 1.
Projecting equation (9) on the (O;X1) gives the
following equation:
them. For the second, third and fourth crankpins, the
N − XB1 (12) expressions are the following:
π π π
About the projection of equation (9) on the (O;Y1) we FA ⎛⎝ θ + ---⎞⎠ , FA ⎛⎝ θ + 2 ---⎞⎠ , FA ⎛⎝ θ + 3 ---⎞⎠ (16)
2 2 2
obtain the following expressions depending upon whether
the contact between the jacket and the piston occurs in push Figure 5 represents the evolution of the force exerted by
(N→ N = N) or opposite side (N<0→ N = N): the connecting rod (b1) on crankpin 1 according to
· crankshaft rotation angle θ.
Y 1- = .y··
– YB1 – P ± f -------
· N mp 1 (13) During the admission and compression process (between
Y1 0 and 360o) and evacuation (between 540o and 720o) the
The sets of equation (6,7,8,12,13,13) allow to determine component of the force exerted by the connecting rod (b1)
the followings five liaison forces N, XB1, YB1 XA1 and YA1 on the crankpin 1 (YA1) varies sinusoid ally. While in the
according to the gas pressure P, the coefficient of the process of combustion and expansion (between 360o and
friction between the piston and cylinder f, and the 540o) it varies non-harmonic. Regarding to the component
crankshaft rotation angle θ. Expressions of these forces are (XA1), the appearance is nearly sinusoidal.
given in appendix A.
3.2.4. Study of engine torque
3.2.3. Case of the connecting (b2) By definition, the engine torque is the product of force
The calculation scheme to determine the forces exerted the applied perpendicular to the crankpin by the connecting rod
crankpin on the connecting rod (b2) is identical to Figure 4 on the crankpin by the radius R = Piston stricke/2.
with different angles: (θ+π/4) instead of (θ+π/4). The Considering the connecting rod (b1) in reference related to
expression of forces are similar to those obtained for the the axis of its cylinder and associating to a reference to the
connecting rod (b1). Indeed ϕ·· 1 and ϕ· 1 are replaced by ϕ·· 2 crankpin (Figure 6), we calculate the torque created by the
and ϕ· 2 . Their expression becomes respectively: connecting rod (b1) on the crankpin1as following (Figure 6):
π
λω sin ⎛⎝ θ + ---⎞⎠
2

4
ϕ·· 2 = --------------------------------------------------
3⁄2
- ( 1 – λ2 ) (14)
π
1 – λ sin ⎛⎝ θ + ---⎞⎠
2 2

4
and
π
– λω .cos ⎛⎝ θ + ---⎞⎠
4
·ϕ = -------------------------------------------
2 - (15)
2 2⎛ π ⎞
1 – λ sin ⎝ θ + ---⎠
4

The forces on the piston of the connecting rod (b2)

denoted FA2 of components XA2 and YA2 are calculated taking
into account equation (14 and 15). As the engine is an four
crankpin engine, the forces on the other crankpins are Figure 6. Calculation scheme the engine torque created by
calculated taking into account the phase difference between connecting rod (b1).
 MODEL FOR DYNAMIC BEHAVIOR OF THE CRANKSHAFT OF AN AIR COOLED DIESEL ENGINE 827

cylinders that causes significant efforts on crankpins.

During the evacuation process, the torque becomes low.
The variation of the engine torque as a function of the
crankshaft rotation angle on the different crankpins and
variation of the resulting engine torque (superposition on
each crankpin torque) are shown in Figure 8 (a and b).

4. MODELING OF BEARING –CRANKSHAFT

In this section, we determine the reactions of bearing on the

Figure 7. Engine torque created by the first. crankshaft. In our case, the hydrodynamic bearing model is
used. A smooth bearing consists of a shaft which rotates
within a bearing bush separated to the latter by a film of
π π
T engine = R ⎛⎝ XA1 cos ⎛⎝ θ – ---⎞⎠ + YA1 sin ⎛⎝ θ – ---⎞⎠ ⎞⎠ (17) fluid. Under the action of a load, the center of the bearing
4 4
bush and the shaft do not coincide, and there is a
By analogy the engine torque created by connecting rod convergent-divergent corner in the film. The rotation of the
(b2) on crankpin 1 is: shaft drives the fluid in the converging wedge and creates a
pressure field that opposes to the load and balances it. This
π π
T engine = R ⎛⎝ XA2 cos ⎛⎝ θ + ---⎞⎠ + YA2 sin ⎛⎝ θ + ---⎞⎠ ⎞⎠ (18) model focuses on the hydrodynamic regime and assumes
4 4
that the fluid flow through the bearing is isothermal and
Finally, the total engine torque on the crankpin 1 is at laminar. Figure 9 shows the schemes for calculating the
each instant equal to the sum of the torques generated by hydrodynamic pressure force of oil film. According to the
the two connecting rods. The torque on other crankpins is assumptions shown in Figure 9 and for the incompressible
determined in the identical way, by taking into account the
phase difference between two crankpins (equation (16)).
Figure 7 shows that during the admission and the compression
process (between 0o and 360o), the engine torque produced
by the first cylinder is relatively low. While the expansion
process corresponds to higher values of the torque. This
could be due to the high pressure gas combustion in the

Figure 8. Variation with the crankshaft angle: (a) of the Figure 9. Scheme and assumptions of computation of the
engine torque on the various crankpins; (b) of the total hydrodynamic pressure of the oil film exerted on the
engine torque resulting from the superposition of the crankshaft: (a) hydrodynamic bearing; (b) bearing bush
torques created on the four crankpins. assimilate to a plan (Jean, 1995).
828 O. KEITA, T. HEDHLI and J. BESSROUR

and iso-viscosious fluid permanent regime, Reynolds Table 1. Crankshaft mechanical properties.
equation is given (Jean, 1995) as:
E Yield
Material ν[-] ρ (k/m3)
∂- ⎛ 3 ∂P ∂ ∂P (N/mm²) (Mpa)
h ------⎞ + ----- ⎛ h3 ------⎞ = ---------- ------
1- ----- 6µV dh
---- (19)
R2 ∂θ ⎝ ∂θ ⎠ ∂z ⎝ ∂z ⎠ R dθ 42CD4 200.000 0.3 7800 780
Where P, is the pressure in the lubricant, θ and Z, are the
circumferential and axial coordinates, respectively, and
expressed as θ = X/R h, is the lubricant film thickness
V = ωR, is the linear velocity of the shaft µ, is the dynamic
viscosity of fluid. ε, is the relative eccentricity (ε = e/C),
which varies between 0 and 1. The film thickness can be
written as:

h = C ( 1 + εcosθ ) (20)

Where C, is radial gap and expressed as: C=R2 – R1.

Considered the boundary conditions of Stierber and
Swift that are, also referred to Reynolds conditions, in the
case of infinitely short bearing, and taking into account the
boundary condition of Gambel that assumes the pressure
zone is limited to the convergent part of the film (Jean,
1995). The pressure field can be expressed as follows:
2
3µV- ⎛ ---- ε sin θ -
L-⎞ -----------------------------
P ( θ ) = --------- (21)
C 2 R ⎝ 4 ⎠ ( 1 + ε cos θ ) 3
where L, is the length of the bearing (Figure 9). C = 0;
1mm. The value of ε is obtained from the abacuses Figure 10. Crankshaft modeling: (a) Mesh; (b) mechanical
according to the ratio L / D of length to the diameter of load applied on crankshaft.
bearing and the dimensionless value of the constant load
acting on the bearing (Jean, 1995) S, defined as the
following equation: the charged surfaces are interchanged in two consecutive
2 time-cycle engines. The surface distribution of efforts has
µLD R
S = ----------- ⎛⎝ ----⎞⎠ (22) been used. On the crankpins, the loads are the reaction
W C
forces of the crankpins against the connecting rods
where W, is the constant load on the bearing. In our case, S calculated from the dynamic characterization as proposed
is the weight of the crankshaft supported by a bearing. in section 3. (ii) -The inertia forces of rotation applied in the
entire volume of the crankshaft. (iii)- The resistive torque
5. MODELING AND SIMULATION OF THE on the flywheel. (iv) -The hydrodynamic pressure forces of
CRANKSHAFT DYNAMIC BEHAVIOR oil film resulting from bearing -crankshaft boundary
conditions. Figure 10 represents the mesh and applied loads
In this section, we present a numerical investigation of the on the crankshaft. For the dynamic calculation, we use the
dynamic behavior of the crankshaft by the finite element dynamics implicit scheme. The mechanical properties of the
method with numerical code ABAQUS. crankshaft are summarized in Table 1.

5.1. Mesh 5.3. Numerical Results

The geometric model has been developed on CAO Among the numerical results from the simulations, we
software and imported into the finite element analysis extract essentially the von Mises equivalent stress. These
code. The mesh is adopted in linear by using tetrahedral results allow us to analyze the fatigue under cyclic loading
element types (C3D4) and, contains 36.215 nodes and (by using the von Mises equivalent stress). Figure 11 shows
149.062 elements. the state of the von Mises equivalent stress for some
instants of the cycle. It proves that high stress zones are at
5.2. Crankshaft Loading Modelization the crankpin 3. This may be due to the fact that the
(i)-The efforts on the crankpins which the correct crankpin 3 occupies the central position of the crankshaft,
application is an important point. These efforts are modeled more solicited by bending stresses. The fatigue analysis of
by some efforts representative at nodes for which it is the crankshaft is performed at this crankpin. The Figure 12
necessary to make certain distribution assumptions. In fact, shows the evolution over time of the most severely stressed
 MODEL FOR DYNAMIC BEHAVIOR OF THE CRANKSHAFT OF AN AIR COOLED DIESEL ENGINE 829

Figure 12. Evolution over time of σMises at the most stressed

point of crankpin.

strength of the material: this is the fatigue phenomenon.

This situation leads to premature appearance of damage
(cracking or fracture).
The fatigue analysis are a group of stresses and strains
post treatment analyses oriented towards to an appreciation
of the safety factor, which reflects the durability or
endurance. For the fatigue analysis, we use the following
fatigue criterion: Haigh and Dang Van criterion.

5.4.1. Fatigue verification with Haigh criterion

Endurance diagram of abscissa, mean stress denoted σa and
ordinate alternating stress denoted σm are from the Wöhler
curves (Figures 13). It defines the set of endurance limits
σD depending on the value of the average stress for a given
number of cycles (Jean, 1995). For the fatigue analysis and
to define a safety factor of mechanical parts on can use the
Haigh diagram. This diagram represents the amplitude of
alternating stress depending on the mean stress. The
construction of diagram can be performed in two ways,
Figure 11. Von Mises stress for some times of the cycle. either from the Goodman line (used in our model) or from
the Gerber parable. In both cases, the factor of safety with
respect to the fatigue limit or endurance is determined
point of the crankpin 3. We can observe fluctuations, as it is using the following equations (Figure 13):
the case of any part submitted to various stresses over time.
In practice for the calculation, these fluctuations are
sinusoidally modelized. The maximum value of the stress
in this point is 46 MPa (Figure 12), which remains
significantly lower than the yield stress of the crankshaft
material: Re = 870 MPa.

5.4. Verification in Fatigue

As indicated in Section 1, the aim of this section is to
identify the fatigue behavior of the crankshaft under
maximum engine loading conditions (N = 2650 tr/mn). The
crankshaft during its motion is subjected to dynamic and
cyclic loads. These solicitations induce cyclical stresses. It
follows that the breakage of the part occurs after a number
of cycles N for a stress amplitude below the static tensile Figure 13. Endurance diagram (Gazaud et al., 1969).
830 O. KEITA, T. HEDHLI and J. BESSROUR

σ
( C, D ) : y = -----D- x + σ D (23)
Rm
( D, E ) : y = – x + R e (24)
σ
( D, A ) : y = -----a- x (25)
σm
Point B is the intersection of the line (O, A) with line of
the Goodman (C, D). Point A can occupy different
positions on the line (O, A). The safety coefficient is
defined as:
OB
S = -------- (26)
OA
The choice of the stress fatigue limit σD, breaking
strength σm and yield strength Re is made according to the
steel grade of the crankshaft, number of cycles, and its
diameter (Jean-Pierre, 2008). For our crankshaft of 42CD4
grade, corresponding value are: σD = 600 MPa, for 107cycles;
σm = 1200 MPa, and Re = 870 MPa. The determination of the
equation of the line (O,A) requires knowledge of mean and
alternating stress (equation (25)). For that we use the Figure 14. Haigh diagram: (a) at crankpin 2, (b) at crankpin 3.
evolution curve of von Mises equivalent stress (Figure 12).
The development at first order in Fourier series of von
Mises stress is expressed as:

σ eq ( x, y, z, t ) = ( ( σ eq ) + ( σ eq )a sin ( ω t + ϕ ) ) (27)

where,
T
( σ eq )m = ∫ σ eq ( x, y, z, t ) dt (28)
0

and
2 1⁄2
⎛⎛ T ⎞ ⎞
⎜ ⎜ σ eq ( x, y, z, y ) cos ω t dt⎟ +⎟
⎜ ∫ ⎠ ⎟
2 ⎝0 Figure 15. Fatigue verification at the crankpin 2. Dang van
( σ eq )a = --- ⎜⎜ ⎟
2 ⎟
(29)
T ⎛T ⎞ ⎟ Diagram.
⎜
⎜ ⎜ ∫ σ eq ( x, y, z, y ) sin ω t dt⎟ ⎟
⎝ 0⎝ ⎠ ⎠

are the mean and alternating stress, respectively. We use the approach. It proposes a linear relationship between the
graphical method of trapezoids to compute integrals in amplitude τa(t) (alternating part) of octahedral shear stress
equation (28 and 29). For the maximum power of regime, and hydrostatic pressure pH(t), and is expressed as follows
the cycles number is: 1cycle = 4π = 12.56rd. For N regime, (Bouraoui et al., 2007):
it corresponds to n cycles /s. n=N/2*60 = 22.08 cycles/s for
maximum power regime, where the pulsation ω = 138.66rd. max ( max ) { [ τa ( t ) ] + αD p H ( t ) })) ≤ βD (30)
n t
Our factor of safety research is focus on a number of points
where the equivalent stress is maximum at a given time in τa(t) is define by the following equation:
the cycle. Figure 14 shows the fatigue verification with
Haigh criterion at the crankpin 2 and 3. τa ( t ) = τa ( t ) – τ m (31)

5.4.2. Fatigue verification with Dang Van criterion where τ(t) and τm are macroscopic and average split,
The Dang Van criterion is one of the multi-axial fatigue respectively. τ(t) can be expressed as:
criteria. A multiaxial fatigue criterion indicates whether the
1
endurance limit or more generally the fatigue limit of the τ( t ) = --- ( σ 1 – σ 2 ) 2 + ( σ 2 – σ 3 ) 2 + ( σ 3 – σ 1 ) 2 (32)
3
material in N cycles is reached for a succession of stress
states defining a multi-axial cycle. This criterion is a global The constants αD and βD are determined from reference
 MODEL FOR DYNAMIC BEHAVIOR OF THE CRANKSHAFT OF AN AIR COOLED DIESEL ENGINE 831

Figure 17. (a) Experimental device and test underway, (b)

frequency spectrum of the Fourier transform of the
response to the excitation.

The measurements only last a few seconds. The

mathematical representation of an ideal shock is realized
by the infinitely short DIRAC function of which amplitude
spectrum is constant over the entire frequency range. The
experimental device consists of a piezoelectric sensor
Figure 16. Representation of free crankshaft eigen modes (accelerometer) which captures vibration signal on the
without engine flywheel for the first sixth modes. crankshaft excited in impulsive mode. The signal from the
sensor is transformed by a signal conditioner which
comprises: (i) a measuring transducer which converts
fatigue tests (Gazaud et al., 1969). (see details in Appendix signal into electric current easily transportable, serves as an
A). The fatigue resistance condition is defined as: input and is powered by a 12 volt source. (ii) an amplifier is
designed to increase the signal intensity and to make it less
τa + 0, 198 p H max ≤ 294, 3 (33) sensitive to the noise of ambient background. It serves as
an output signal, and it is connected to the Fourier
The Figure 15 illustrates Dang van diagrams at transformer. A source of current supply of 12 volts is
thecrankpin 2 connected to the converter. A spectrum analyser, fast
Fourier transform, gives the Fourier transform of temporal
6. MODAL ANALYSIS OF THE CRANKSHAFT signal. Figure 17 (a) shows the experimental device used for
the completion of vibration test, and Figure 17. b
6.1. Numerical Study demonstrates the frequency spectrum of the Fourier
Vibration analysis of the first six modes of the free transform response.
crankshaft (without flywheel) is shown in Figure 16. The
first vibration mode (f = 4.11 Hz) corresponds to bending 6.2.1. Measurement results
in the vertical plane with two nodes of vibration which are After repeated the test several times on a frequency band of
relatively symmetrical to the crankshaft center. The second 50 Hz, the first eigen frequencies obtained with the free
vibration mode (f = 4.23 Hz) is a bending mode in the crankshaft without engine fywheel are: 4.6 Hz; 8.7 Hz;
horizontal plane also with two vibration nodes relatively 11.8 Hz; 17.8 Hz. and 21.4 Hz. Table 2 shows the comparison
symmetrical to the crankshaft center. The fourth vibration between the first five eigen frequencies determined in the test
mode (f = 9.87 Hz) is a torsional mode along the axis of the and the values computed by numerical simulations.
crankshaft with a single node near by the engine flywheel:
is the first torsional mode. The fifth vibration mode 6.3. Interpretation of Test Results
(f = 11.49 Hz) is a bending mode in the vertical plane with The relative error between the eigen frequencies for
three vibration nodes. experimental and numerical values, varies from five (8 −

6.2. Experimental Study Table 2. Comparison of experimental and numerical eigen

The test is performed in the laboratory of mechanical frequencies.
vibration of Ecole Nationale d'ingénieurs de Tunis (ENIT).
The aim of this test is to compare the free crankshaft M1 M2 M3 M4 M5
frequencies with the numerical simulation results. We used (HZ) (HZ) (HZ) (HZ) (HZ)
the impulse shock close to a Dirac due to its ease of Experimental frequencies 4.6 8.6 11.8 17.8 21.4
production and rapidity. We use the averaging technique of
Mumerical frequencies 4.11 8.9 9.87 18.1 23.07
a predetermined number of spectra to eliminate the noise.
832 O. KEITA, T. HEDHLI and J. BESSROUR

16.35) %. These errors could be due the one hand, to the (1969). La Fatigue des Metéaux. DUNOD. Paris.
fact that the geometric model is an idealized geometry of Heath, A. R. and McNamara, P. M. (2008). Crankshaft
the crankshaft, not taking into account all its real form, and stress analysis - The combination of finites element and
secondly, they may come from: (i) systematic error; a classical techniques. J. Eng. Gas Turbines Power 112, 3,
typical example is the drift time of the sensor sensitivity. 268−275.
(ii) the parasite error; it is defined in standard NF X 07-001 Henry, J. P., Toplosky, J. and Abramczuk, M. (1992).
as a gross error resulting from improper execution of Crankshaft durability prediction a-new 3D approach.
measurement. Table 2. Comparison of experimental and SAE Paper No. 920087.
numerical eigen frequencies. Ide, S., Uchida, T., Ozawa, K. and Izawa, K. (1990).
Improvement of engine sound quality through a new
7. CONCLUSION flywheel system exibly mounted to the crankshaft. SAE
Paper No. 900391.
This paper has enabled the development a numerical model Jean, F. (1995). Butées et Paliers hydrodynamique.
of three dimensional global dynamic behavior of the Techniques de L'ingénieur. B-5320.
crankshaft by finite elements of a diesel engine air cooled. Jean-Pierre, P. (2008). Moteur Diesel D'automobiles,
To achieve this goal, we carried out a kinematic and Conception et Mise en Point. Techniques de L'ingérieur.
dynamic analysis to determine the forces and moments on BM-2575.
different crankpin of the crankshaft. Description of the Jian, L. and Henri-Paul, L. (1998). Concentration de
stress fields of mechanical origin, led to the identification Contraintes. Techniques de L'ingénieur. BM-5040.
of t most solicited zones which are the risk zones. To Kim, M. S. and Woo, Y. H. (2013). Robust design
validate the dynamic endurance, a analysis of the fatigue optimization of the dynamic responses of a tracked
behavior was carried out in these zones. That, using Haigh vehicle system. Int. J. Automotive Technology 14, 1, 47−
criterion and also the multiaxial fatigue criterion of Dang 51.
Van that takes into account the multiaxial cyclic loading of Larsson, S. and Schagerberg, S. (2004). SI-engine cylinder
crankshaft. Simulation of vibration behavior by finite pressure estimation using torque sensors. SAE Paper No.
elements of free crankshaft allowed identifying eigen 2004-01-1369.
modes and frequencies. An experimental vibration test Nikolic, N., Torovic, T. and Antonic, Z. (2012). A
conducted on free crankshaft able to validate the model by procedure for constructing a theoretical wear diagram of
comparing the eigen frequencies experimentally obtained IC engine crankshaft main bearings. Mechanism and
to those found in numerical simulation. Comparison of Machine Theory, 58, 120−136.
values is relatively good. Niinoa, T., Iwamoto, T. and Ueda, S. (2002). Development
of simulation technology for dynamic behavior of
ACKNOWLEDGEMENT−We thank Islamic developpement crankshaft systems in motorcycle engines. JSAE Review,
Bank (IDB) scholarship program for financially support this 23, 127−131.
work. Schagerberg, S. and McKelvey, T. (2003). Instantaneous
crankshaft torque measurements - Modeling and
REFERENCES validation. SAE Paper No. 2009-01-0713.
Swoboda, B. (1998). Mécanique des Moteurs Alternatifs.
Bouraoui, C., Fathallah, R., Khantouch, Z., Châteauneuf, Institut Française du Pétrole. Edn. Technip. Paris.
A. and Destrebeq, J. F. (2007). Analyse comparative des Taher, H. (2003). Un Modèle Thermomécanique d'un
critères de fatigue multiaxiale basée sur une approche Moteur à Combustion Interne Refroidi Par Air. Mémoire
fiabiliste: Etude de cas. Congrès CMSM 2007. Monastir, de DEA. ENIT.
Tunisie. Taraza, D. (2002). Accuracy limits of IMEP determination
Çevik, G. and Gürbüz, R. (2012). Evaluation of fatigue from crankshaft speed measurements. SAE Paper No.
performance of a fillet rolled diesel engine crankshaft. 2002-01-0331.
Engineering Faliure Analysis, 27, 250−061. http:// Yu, S., Choi, H., Cho, S., Han, K. and Min, K. (2013).
dx.doi.org/10. 1016/j.engfailanal.2012.07.026 Development of engine control using the in-cylinder
Dimbale, N., Bhujbal, A., Hanabar, R., Gawande, S. H., pressure signal in a high speed direct diesel engine. Int.
Navale, L. G. and Daphal, G. (2011). Dynamic vibration J. Automotive Technology 14, 2, 175−182.
analysis for multi-cylinder diesel engine of SL90 type.
Int. J. Science and Advanced Technology 1, 2, 55−61. APPENDIX
Feilong, L., Gehan, A. J., Nick, C. and Ahmed, S. (2012).
An experimental study on engine dynamics model based A. Kinematics of the Piston and the Connecting Rod
in-cylinder pressure estimation. SAE Paper No. 2012- Appendix A.1. Kinematics of the connecting rod
01-0896. The velocity field of a solid being a equiprojectif vector
Gazaud, R., Pomey, G., Rabbe, P. and Janssen, C. H. field, we can write:
 MODEL FOR DYNAMIC BEHAVIOR OF THE CRANKSHAFT OF AN AIR COOLED DIESEL ENGINE 833

v ( G ) = v ( A ) + Ω^AG (A1) (equation (6)) as defined in section 3

Where, Appendix A.3. Determination of the Unknown Liaison

Ω = ϕ· ⋅ z (A2) Forces
By solving the sets of equations (6,7,8,12,13,13) in section
is, rotation vector of the connecting rod v(G) is, velocity of 3 we obtain:
the gravity center G and v(A) is, Velocity of point A.
However, the velocity of point A linked to the connecting XB1 = N (A8)
rod is equal to the velocity of point A linked to the crankpin
(A being the geometrical point of connecting rod-crankpin For the case of N > 0 ( θ ∈ [ 0, 2π ] ∪ [ 2π, 3π ] )
contact, see diagram), whence: YB1 is obtain in the following expression:
y· 1
v ( A ) = – ( ωr cosθ ) . x – ( ω r sin θ ) . y (A3) YB1 = – mp .y··1 – P – f ------ N (A9)
y· 1
knowing that: Substitution of equation (A.9) into equation (9) (in
section 3) lead to the following expression:
AG = l G sinϕx + lG cos ϕ.y (A4)

Equation (A.1) lead to: (A10)

vG = – ( ωr.cosθ + ϕ· lG cos ϕ ).x – ( ω r sin θ + ϕ· lG sin ϕ ).y (A5)

For the case of N<0 ( θ ∈ [ 0 – π, 2π ] ∪ [ 3π, 4π ] ) similarly
By deriving equation (A.5) with respect to time and we find YB1 as:
remembering that ω is considered constant, we obtain
y· 1
equation (5) as mentioned in section 3 YB1 = – mp ⋅ y··1 – P – f ------ N (A11)
y· 1
Appendix A.2. Kinematic of the Piston In this second case N is expressed as:
Projecting the contour of OAB on the axis (O; y) and
taking into account equation (1), we obtained elongation of
the piston as follows: (A12)

y = r cosθ + l ( 1 – λ2 sin2 θ ) (A6)

The components of the force exerted by the crankpin on
By deriving equation (A.6) with respect to time one the connecting rod (b1) XA1 and YA1 can then be determine
obtains the velocity of the piston as follows: by following below expressions:
cos θ
y = –ω rsinθ 1 + ------------------------------------- (A7) XA1 = m b1 x··G1 – XB1 (A13)
( 1 – λ 2 sin 2 θ )
Deriving a second time equation (A.6) and taking account YA1 m b1 y··G1 – YB1 (A14)
equation (1), we obtain the acceleration of the piston
Appendix A.4. Determination of the Constants αD and
βD for Dan Van Fatigue Criterion

Constants αD and βD are determined from reference

fatigue tests and expressed as:
t –1 – f –1 ⁄ 2
αD = ---------------------
- (A15)
f –1 ⁄ 3
β D = t –1 (A16)

t-1 and f-1 are the endurance limits in torsion and alternating
bending, respectively. For the material of the crankshaft
42CD4 : t-1 = 294.3 MPa, f-1 = 519.93 MPa. This allows to
Figure A18. velocity scheme. find αD =0.198 and βD =294.3.

View publication stats