Elsevier

Composite Structures

Volume 89, Issue 2, June 2009, Pages 285-293
Composite Structures

Influence of stress relaxation on clamp-up force in hybrid composite-to-metal bolted joints

https://doi.org/10.1016/j.compstruct.2008.07.031Get rights and content

Abstract

This paper presents an experimental investigation of the influence of stress relaxation on clamp-up load in hybrid composite-to-metal bolted connections. Loss of bolt clamp-up load may influence the strength and fatigue life of the connection. This study focuses on the effects of bolt retightening, use of tapered-head bolts versus protruding-head bolts, and briefly addresses environmental effects including temperature. All tests were conducted for a time period of at least three months in order to estimate primary and secondary stress relaxation effects. Bolt clamp-up load curves were fit to three different but similar equations for stress relaxation using a least squares method. The bolt retightening tests show that some of the preload in the connections can be maintained with periodic retightening of the bolts. Little difference in relaxation was observed when comparing tapered-head bolts with protruding-head bolts, given that roughly equal stress relaxation rates were observed. In a limited amount of testing, it was observed that temperature shifts caused more rapid stress relaxation rates and that post cure of the composite is essential to achieve a stable thermal response.

Introduction

The present experimental study of stress relaxation in hybrid composite/metal bolted connections is part of an ongoing effort to investigate the response of hybrid composite/metal connections used in US Navy vessels. Bolted connections are used in structures where the ability to easily remove structural components is required. Concern over the effect of stress relaxation due to the viscoelastic nature of the composite material was the impetus for this study. Stress relaxation due to creep in the through-the-thickness direction of the composite material typically occurs in bolted composite connections. Accordingly, a composite panel bolted in the thickness direction is susceptible to loss of clamp-up load, due to the viscoelastic nature of the resin, which dominates the through-the-thickness direction response. Loss of clamp-up load may have a deleterious effect on the strength and the fatigue life of the connection.

The US Navy currently has interest in developing hull-forms using advanced materials such as composites and corrosion resistant metals in order to enhance its future naval capabilities. Meeting stiffness requirements in some applications may be difficult with E-glass/vinyl ester (EG/VE) composites due to their low material stiffness when compared to a metal such as steel. Hybrid construction has merged the use of metals and composites as a way of implementing EG/VE systems into the bulk of the structure, while maintaining adequate stiffness. Barsoum [1] describes the development of hybrid hulls where non-magnetic stainless steel is used in combination with EG/VE composites to create advanced hull-forms with low electromagnetic signatures. He stated that EG/VE composites alone lack the stiffness and in-plane strength required for the larger ship hull structures. Alm [2] estimated that a 50-m composite naval ship would be 2.4 times less stiff than its steel counterpart indicating the potential benefit of hybrid systems. Kabche et al. [3] described a hybrid system where composite panels are attached to a metallic sub-structure; which provided the impetus for this research effort.

Oftentimes it is feasible to attach dissimilar materials using bonded connections to reduce the effects of stress concentrations. In some instances, however the use of bolted hybrid connections is necessary, especially in cases where removable panels are required. The hybrid joint presents a greater challenge than that found in joining similar materials, mainly due to the differences in the constituent material properties. While bolted connections offer easily removable parts, the bolt clamp-up load may be compromised by bolt stress relaxation. Composites tend to creep over time due to the viscoelastic nature of the matrix material, which leads to potential loss of bolt clamp-up load. In marine structures, this can adversely affect the watertight integrity of the joint. In order to maintain a watertight connection, stress relaxation in the bolts must be adequately understood. Relaxation effects need be quantified so that they are properly accounted for in design. The goal of this study is to perform a preliminary assessment of the important parameters that govern stress relaxation in hybrid composite-to-metal bolted connections. In this way, watertight, cost effective and reliable connections may be developed.

Maintaining adequate bolt clamp-up load can significantly affect the strength of the connection. One example is shown by Sun et al. [4] who demonstrated the importance of clamp-up force on bolt strength through finite element analysis and experimental validation. The strength of double lap shear pin joints made of T800H/3900-2 graphite epoxy was highly dependent on clamp-up stress. Their work also showed that increasing the clamping area resulted in stronger joints. Another example is the work by Starikov and Schon [5] on fatigue of carbon fiber/epoxy bolted joints loaded in-plane. Their research demonstrated that the clamp-up load can influence the fatigue life and also that the applied load influences the clamp-up stress distribution as the connection deforms.

Several methods have been used over the years to model the creep behavior of materials. These methods are formally described by Findley et al. [6]. In composite materials, the creep effects are due primarily to the viscoelastic properties of the resin. Creep of polymer composites is a non-linear, time-dependent phenomenon and is affected primarily by stress history, temperature, environment and time [7]. A study by Shen et al. [8] found that creep was extremely sensitive to water; the immersion of composites in water accelerated the recovery of the surface due to creep. Other factors affecting the creep response of a composite system are the type and architecture of the reinforcing fibers, temperature, and humidity. For a given strain rate, the relaxation behavior of the composite is very similar for both loading and unloading, according to Kim and Sun [9].

Much past research has been done on creep response in the fiber direction, where creep is dependent upon the properties of the matrix material, the fiber material and the fiber volume fraction. Kim and McMeeking [10] performed tests on creep in composite materials in the fiber direction and showed that fiber slipping and/or mass transfer can greatly reduce the creep strength of a composite system. Raghavan and Meshii [7] performed studies on the long-term deformation and strength of composites using short-term test data obtained for accelerated testing conditions such as higher temperature, stress, and humidity. Maksimov and Plume [11] researched the effects of different fiber materials on the creep of the composite. Scott and Zureick [12] performed long-term testing on pultruded E-glass/vinyl ester composites under longitudinal compressive loading and modeled their experimental results using the power law developed by Findley et al. [6].

The creep response in the direction perpendicular to the fibers (through-thickness direction) is almost entirely dictated by the matrix material in a bi-axially reinforced laminate. In bolted connections, this gives rise to the stress relaxation phenomenon, where creep effects are typically observed in the through-thickness direction. Hence, in time, a composite bolted in the thickness direction is highly susceptible to preload loss, as demonstrated by Weerth and Ortloff [13]. According to Pang and Wang [14], adding stitching in the through the thickness direction of a composite can improve the overall creep resistance of bolted joints since the bolt load causes stress in that direction. Experiments and numerical modeling performed by Chen and Kung [15] have shown that the preload in bolted joints is highly sensitive to changes in temperature and humidity. As observed by Guedes et al. [16], the time-dependent properties of the composite are most largely influenced by the resin and the interface between the fiber and the resin. Environmental factors, such as humidity, are typically coupled with the mechanical loading of the composite joint and the mechanical response of the composite is dependent upon exposure time.

Shivakumar and Crews [17] presented their work on bolt clamp-up relaxation in simple T300/5208 graphite/epoxy connections. They concluded that clamp-up force undergoes significant relaxation even at room temperature dry conditions. Relaxation of 30% was predicted for exposure duration of 20 years. Increased rates of relaxation were shown at elevated temperature and moisture content. They performed viscoelastic finite element analysis with some experimental verification. This work included the effects of temperature and humidity and they developed the expression relating the non-dimensional clamp-up load loss versus time as follows:PtP0=11+F1αTHn·tnwhere Pt is the load at time t, P0 is the initial clamp-up load, F1 is a material dependent constant, n is the viscoelastic power law constant for the material and αTH is a hygrothermal shift factor which accounts for variations in temperature and humidity. When no hygrothermal shift is considered this expression can be reduced to the following:PtP0=11+K1·tnFox [18] presented a study of bolted E-glass/vinylester joints using 15.9 mm (5/8 in.) diameter load sensing bolts with a tapered-head adapter to quantify the effect of tapered-head bolts on the connection response. The study used a lap shear connection geometry and quasi-isotropic layup for the composite bolted to a 9.5 mm (3/8 in.) thick steel plate with washers below the plate. Both relaxation and lap shear tests were performed. Relaxation tests were carried out for approximately 120 h. The data reduction and least squares fit was performed using Eq. (2). For highly torqued tapered-head bolts, the constants were found to be K1 = 0.0861 and n = 0.2519. It was also found that if the initial clamping force is low, load loss can decay to near zero in a very short period of time. The data on bearing strength tests suggested that there may be a moderate increase in strength with clamp-up load.

Weerth and Ortloff [13] studied composites in a bolted connection using a test set up that consisted of an E-glass/vinyl-ester cylindrical coupon sandwiched between 2 washers held together by a bolt. The load history was read through the use of washer load cells and a computerized data acquisition system. A 15.9 mm (0.625 in.) diameter bolt was used in conjunction with a composite having a bearing area of 298 mm2(0.462 in.2). Specimens were preloaded at an ambient temperature of 21 °C (70 °F) which was increased to 49 °C (120 °F) during the relaxation phase. Tests were run for periods of 1 day, 1 month, 1 year, 5 years and 10 years. The loss of preload was found to be 15%, 28%, 36%, 41% and 43%, respectively. The load loss data was found to fit the power law form, and the preload loss was predicted using the following power law equation:Pt=P0·t(1.482×10-11·P02.244-0.0497)where Pt is the load in pounds in the connection at any time, t, and P0 is the initial preload. The constants in the exponent of this expression were computed by curve fit in US customary units to the data collected from six test samples tested using a preload ranging from 56 kN (12,621 lbs) to 1.4 kN (314 lbs) resulting in an average composite material stress range of 188–4.7 MPa.

A simple power law expression can be used be to account for a shift in time required to compute the load at t = 0 as follows:Pt=P0·(1+t)-αThis expression results in the initial preload at t = 0 and includes a single parameter, α, in the curve fit. The value of α may depend upon such variables as temperature, humidity, moisture content, stress and thickness. Use of the single parameter allows for a simple numerical case by case comparison when compared to the 2-parameter model in Eq. (2). A 2-parameter power law expression for computation of the bolt load loss with respect to time was employed prior by Pelletier et al. [19] for bolted hybrid connections. This power law expression is given by Eq. (5) as follows:Pt=β·P0·(1+t)-αIn this expression α and β are constants that depend on constituent materials, geometry and test conditions. This expression normally results in a closer curve fit than Eq. (4) but does not result in the initial preload at t = 0 unless β is equal to unity. A value of β other than one may be due to the continually changing load value at the start of the tests which is affected by the primary creep response.

Section snippets

Test objectives

The objective of the experimental effort described herein is to characterize the stress relaxation of hybrid bolted connections under normal use conditions. Hybrid connections were studied at a sub-component level in order to isolate the effects of the viscoelastic creep on the relaxation of the bolt. Several parameters were studied in this investigation as follows: (1) bolt stress distribution; (2) bolt retightening; (3) varying thickness of the constituents; (4) varying bolt size; (5) type of

Testing method

Four different test procedures were utilized for this experimental investigation to observe the effect of the different parameters that affect the creep response of hybrid joints. These tests were: compression block tests, bolt retightening tests, bolt type tests, and elevated temperature tests. The objectives and procedures for each test are described below. Testing also included a control sample where a bolt was attached to an aluminum plate with no composite in place. A computer controlled

Compression block test results

Fig. 5 shows the load history of the compression block test of two specimens preloaded to 184 and 89 kN. The clamp-up load is non-dimensionalized by dividing the load at time t, Pt, by the initial preload, P0. The actual initial preload on each the four bolts in the 184 kN test was 46.6, 45, 47.8 and 44.5 kN at an average pressure of 71 kPa. The actual preload one each bolt for the 89 kN test was 22.3, 22.3, 22.3 and 22.3 kN at an average pressure of 34 kPa. An average 13% reduction in the initial

Conclusions

Clamp-up load loss in hybrid composite-to-metal bolted connections is significant and highly variable. Causes of the variability include differences in material properties typical of composites, geometry changes, surface finish, workmanship and other effects. Loss of bolt-clamp-up load may influence the strength and fatigue life of the connection and is warrant of further study. The results of the present study showed that after 2000 h of relaxation response clamp-up load loss varied from 13%

Acknowledgements

The authors gratefully acknowledge funding for this project through the Office of Naval Research under grants number N00014-01-1-0916 and N00014-05-1-0735. Dr. Roshdy G. S. Barsoum of ONR is the cognizant program officer. Guidance provided by Drs. Milt Critchfield, Loc Nyugen and Gene Camponeschi of NSWC-Carderock is gratefully acknowledged.

References (19)

There are more references available in the full text version of this article.

Cited by (56)

  • Preload relaxation analysis and reliable life prediction of space connection and separation device based on accelerated degradation tests

    2023, Chinese Journal of Aeronautics
    Citation Excerpt :

    The gap adjustment and deformation coordination within the structure cannot be reasonably considered by physical analysis approaches. Test analysis methods, from another perspective, investigate the preload relaxation law by conducting tests and derive empirical preload formula based on measurements.12–15 They are suitable for situations focusing on the preload relaxation analysis of devices in practical engineering because these approaches are often more intuitive, and require fewer model parameters compared to physical analysis methods.

View all citing articles on Scopus
View full text