lucraresipa'09

8/8/2019 lucrareSIPA'09

1/4

ON THE FRETTING-FATIGUE BEHAVIOURAT THE STEEL BOLTED ASSEMBLY

Mircea VODA*, *, Eugen S. ZABAVA*, *,

* Politehnica University of Timisoara, Mechanical Faculty

e-mail: [email protected], , [email protected]

Abstract. Aluminium alloys and steel recently are used in food industry. But their mechanical characteristics do not

allow for the complete replacement of steel, more resistant to shock level. The design of such assemblies must to

have in mind an optimum combination of mechanical properties of materials and naturally the understanding of

physical phenomena governing the damage and the reliability of the structures. The fretting fatigue tests were

carried out on an aluminium-stainless-steel assembly. The work has therefore focused on understanding the

phenomena of degradation on bolted joints specimens type bi-materials and on the numerical modeling in order to

describe these tests.

Keywords:fretting fatigue, bolted assembly, riveted assembly, concentrated stress, reliability.

1. INTRODUCTION

The bolted and riveted assemblages are very studied in the present [2, 3], but the number of

parameters and stresses involved in their behavior makes these studies even incomplete. The

assembly constraints of the different parts as such assemblies involve generally significant stressconcentrations in the materials. In effect, in aeronautics, the riveted or bolted assemblies

represent up 95% of junctions and are the most critical parts. The plates holes for assembling are

introduced weakened areas from which fatigue cracks and fretting can emerge and spread.Riveted joints prevent, by their geometry, detection and monitoring of cracks. Duan R. [1] within

the team mechanics and materials IUT A, of Mechanics Laboratory from Lille, proposed in

2006 a device to study the phenomena of fretting and fatigue in joints riveted aluminium and

steel. Our work is focused on the study of this method, the interpretation of results and validation.This research will study real-time initiation and propagation of cracks to a better understanding

and modeling of these phenomena. To anticipate the behavior of such an assembly, it is essential

to differentiate the various stresses to which it will be confronted. There are four: the fretting-wear, the fretting-wear under stress, fatigue and fretting-fatigue.


2/4

2. FRETTING-FATIGUE TESTS

The fretting-fatigue tests were carried out on an assembly aluminium and steel. This case is oneof the closest to reality, especially for an assembly. The specimen was subjected to an axial cyclic

tensile. The specimens consist of two plates, the first aluminium and the other steel bolted with

an special device (see fig.1). The perfect contact between the two plates is only theoretical,

because it is practically impossible to obtain in reality.

2.1. The Assembly used in testing

As aluminium were used a 6061 aluminium alloy, SR EN 573-3:

Tensile (Rm): 315 MPa

Conventional limit of elasticity (R0.2%):196 MPa

Lengthening (A%): 12%

Elasticity modulus (E): 71 GPa

Poisson's ratio (): 0.33As Steel were used a 42CrMo4, EN 10083-1 used in aluminium / steel

assemblies in the field of railways [***:Engineering technology.]

The schematic design of the plates used in specimens is presentedbelow:

Fig.1 Aluminium / steel

assemblage with device

Fig.2 Design of plate test specimens

2.2 The device for obtaining measurements

The measuring device for simulating a riveted assembly with a tubular form, developed by LML

[3], with a tubular form, was provides with Gauges and linked to an acquisition device with

amplification and signal processing in order to measure the local stress around the bolt. Onecalibration was realized in order to place the device in identical conditions to those of a riveted

assembly (see fig.3). Indeed, the gauges are excited by a voltage of 10V, and therefore can not

return a higher signal. But in our previous calibration, a voltage to 10V showed a clamping force

equal to 14.5kN and a rivet assembly subjected to a force of 17kN. We must therefore reduce thegain so that we can reach 17kN while remaining less than 10V.

So:

AE

KUFKKUKeKe MMMf

44===

(1)

With a gain of 700, we obtain, theoretically:

43,04

==

AE

KUK

F

eMf (2)


3/4

Fig.3 Calibration the against-body for a gain of 700

2.3 Tests

For testing, we wanted to stay under the same conditions as Duan R. [2], during testing on

aluminium / steel assemblies. The test conditions are:- Medium load 22 kN;

- Amplitude of the load 4,5 kN;

- Frequency 35 kHz.

Fig.4 Photo of steel and aluminium plates at the end of 2-nd

test, side against-body

The first test was conducted with an 11kN clamp; the against-body had not yet suffered thesecond calibration. The test was stopped at 2,370,700 cycles; no cracks had appeared and test the

most important being that in an effort to higher clamp. The second effort should be 17kN, but the

bolt made available to us does not support the charge. At each attempt, the nets were crushed and

the assembly was loose. Therefore, the load was 13kN and not 17kN as expected. The trial wasstopped at 3,718,430 cycles. On the external side of the test piece of aluminium used in the

second test, we can see the damage due to fretting wear and a cracking start at the bottom of the

image. This crack is due to fretting between the against-body and the piece of aluminium. Theoutside of the steel plate also shows damage due to fretting wear between the bolt and the plate.

As expected, no crack appears on this material. At joining side, the steel and aluminium plates are

only fretting wear due to the connection between the plates. No cracks are visible.

3. STUDY BY FINIT ELEMENT

To get an idea of the results of our experiments, we perform some static FEA analysis. We mustnotify that we have preformed only in the elastic field. The results will be indicative. The

dynamic is the repetition of the static; it is interesting to study each case separately.


4/4

3.1 Aluminium plate

The behavior of an aluminium / steel assembly is unknown; we wanted to study its response topure traction. Knowing the deformation curve as a function of the tensile force applied, we

perform several simulations on the test piece of aluminium considering still tightening effort. The

results correspond to the results logically expected. Nevertheless, this case is far from the

assembly. So do not rely on a simplistic model.

3.2 Steel plate

The analysis on steel plate with various values gives the same localization of the maximal stressas in the aluminium case.

4. CONCLUSION

- Using the measurement device with the gauges it was possible to identify different stages during

the initiation and propagation of cracks which are due to damage by fretting or fatigue on the

specimens.- The use the device as a sensor to determine the initiation of a crack in a test tube to the hole

seems to be very effective.- In the case of assembly the gauges seem to be susceptible to deterioration.- The difficulty to finding a model in FEA close to reality and the impossibility to study the

phenomena of fatigue and fretting does not allow us to anticipate the results obtained

experimentally. However, some similarities are observed.

- Future tests on bolted assemblages would be interesting to get a clear idea of the capabilities ofthis system and confirm the results obtained in this study.

REFERENCES

lucraresipa'09

Documents