Creep Testing
Creep Rupture
Key Observations
References
Creep is the increase in strain or deformation of a material (or structure) with time when the material is subjected to a constant load for an extended period of time (i.e. time-dependent deformation). The change of strain at any time increases with load, temperature, relative humidity and time. Visco-elastic materials, such as adhesives, can undergo creep deformation at relatively low stress levels (well below the ultimate strength of the material) and at low temperatures (i.e. room temperature - referred to as cold flow). This can lead to considerable reduction in life expectancy of the component.
Creep data is often presented as a plot of creep compliance versus time with stress and temperature kept constant (see Figure 1). The creep curve can be divided into three regions. An induction period in which no deformation occurs, a period of logarithmic creep in which creep increases at a relatively uniform rate and finally an accelerated stage terminating in failure by stress rupture. If creep occurs in a bonded joint, there is typically a delay between initial loading and the onset of creep, as clearly shown in Figure 1.
Figure 1: Creep compliance curves for an epoxy adhesive at different levels of stress
Creep tests are performed to assess the extension of joints under load in order to predict long-term behaviour or to assess the long-term strength (creep rupture) of joints under load. High precision extensometry techniques are generally required to monitor joint extension and the tests must be performed under stable environmental conditions (temperature and humidity) to avoid artefacts in the measurement. These tests could, in theory, be performed using any of the loading options outlined below, although the highest accuracy is achieved using either option (a) or (b).
a) Servo-hydraulic test machines;
b) Dead-weight and lever creep testing machines;
c) A screw jack in series with a load cell (Figure 2); and
d) Self-stressing fixture where specimens are placed in either a tube (Figure 2) equipped with a pre-calibrated spring system for loading specimens [2] or a circular ring (Figure 3).
|
|
|
|
Screw-jack test machines |
Self-stressing loading tube |
Figure 2: Methods of loading bonded joints
Figure 3: Self-stressing circular loading rig
The use of a servo-hydraulic test machine is not an economic option in most cases. A bank of small creep machines can be assembled at a considerably lower cost compared with the capital outlay involved with purchasing and operating servo-hydraulic units. Self-stressing fixtures, which are light and economic to produce and maintain, are particularly suited for field trials and for large batch testing. Care should be taken to ensure that the thermal mass of the tubes does not exceed the capacity of the conditioning cabinet, thus preventing correct maintenance of humidity and temperature.
Small single-lap and T-peel joints have been successfully tested using self-stressing tubes. Testing consists of placing specimens in a tube equipped with a pre-calibrated spring system for loading the specimens (BS ISO 14615) [2]. The spring system can be compressed and locked in place to apply the required load with the spring stiffness determining the load range. The amount of load is determined by measurement of the spring compression. The fixture specified in BS ISO 14615 (Figure 2) is capable of loading a series of 3-6 specimens at a time. The specimens are bolted together with either stainless steel or polyamide bolts. The tubes should be suspended vertically within the environmental cabinet to ensure uniform exposure of the test specimens.
The stress tubes are inspected at frequent intervals to check on the condition of the test specimens (i.e. failed or intact). Failed joints are replaced with spacers and the remaining specimens re-stressed. The failure times are measured at which the first three specimens fail. When the third specimen fails, the remaining specimens are removed from the loading tube and tested to failure to determine residual strength. The average lifetime of the failed specimens and the residual strength of the remaining specimens should be recorded. The large uncertainty associated with time-to-failure measurements, especially at the high stress levels will require either electromechanical or optical devices to monitor load or deformation in order to accurately determine time-to-failure.
Specimens loaded by springs can often be in an unstressed state for a considerable period of time (overnight or weekends) before the failed joint is replaced (by a "dummy" specimen) and the loading train is re-tensioned. There is also a tendency for surviving specimens to be damaged in the re-stressing process with the probability of occurrence increasing at high stresses. Creep/relaxation histories of specimens will be different due to the replacement of failed specimens and subsequent re-loading. This contributes further to the uncertainty of creep rupture data. For long term tests over months or years, this effect will probably be minimal.
Load levels need to be established for any particular system tested. Typically these are between 10 and 50% of the short-term strength of the joint. It is generally recommended that that sustained stress in an adhesive bonded joint under service conditions should be kept below 25% of the short-term strength of the joint [1].
For joint characterisation purposes it is recommended that specimens are mechanically loaded at each of five stress levels (i.e. 80%, 70%, 55%, 40% and 25% of the short-term strength of the joint). The large uncertainty associated with creep test results, especially those obtained under hot/wet conditions, implies that the current approach of conducting three tests per stress level is inadequate and that considerably more data points are required for generating reliable creep rupture curves for engineering design purposes. Five (preferably 10) per stress level with five stress levels (see above) per condition should provide a reasonable number of data points.
Failure through creep tends to follow a linear relationship of log (time to failure) to stress applied to the bond, other than at very low stress levels. The gradient of the slope will be dependent on a number of factors, including adhesive type, joint geometry, loading mode, surface treatment and environmental conditions (e.g. temperature and humidity).
Next: Creep Rupture
« BACK TO DESIGN, PREPARATION AND TESTING
