Friction welded steel bars - a fatigue study

TWI Bulletin, January/February 1994

Siak Manteghi is a Principal Research Engineer in TWI's Engineering Department. After gaining a BSc(Hon) in Civil Engineering from UMIST, he continued his postgraduate studies there, working on the stress analysis and fatigue properties of welded joints. He was awarded a PhD for this in 1991.

In 1988 he joined the then Fatigue Department at TWI, and has since maintained his general interest in the structural integrity of welded joints. He has a particular interest in the fatigue performance of non-arc welds, and is currently helping to generate design data for two such processes: friction welding and electron beam welding.

Friction welded components are used increasingly by industry, sometimes in applications which involve fluctuating service loads. Siak Manteghi reports the results of a recent investigation, suggesting that friction welding can be used for joining solid bars when high fatigue resistance is required.

Friction welding is often used as an alternative to arc welding processes to make butt joints between two solid bars, or between a solid bar and another solid component. For instance, in earth moving equipment, friction welding is now an established alternative to more conventional CO ₂ welding, for joining eye-ends to shafts in the hydraulic ram assembly. Friction welded joints of this kind are often subjected to complex and fluctuating service stresses.

Whilst a wealth of information exists on the fatigue behaviour of various weld details produced by MMA and other arc welding processes, very few fatigue data are currently available for joints produced by friction welding. To fill some of this void in our knowledge, fatigue tests have been performed on friction welded joints in a medium strength plain C-steel.

The work reported here is the initial phase of an ongoing project studying fatigue and fracture of non-arc welds. The project is part of TWI's Core Research Programme (CRP) for the period 1992-94, funded jointly by Industrial Members of TWI, and the UK Department of Trade and Industry. The results of the first phase will shortly be published as a TWI Members' report ^[1] .

The studies have since continued, and tests are currently in progress using typical friction welded components fabricated for industrial applications. The results of this more recent work will be reported in due course. The fatigue and fracture properties of joints produced by another non-arc process: electron beam (EB) welding will also be studied as part of the same project.

Experimental

Friction welding

All the specimens were fabricated from 25mm diameter solid bars to BS 970 Grade 080M40 which corresponds closely to the AISI 1039 and 1040 steel grades. This is a medium strength plain C-steel, with yield and tensile strengths of around 750 and 800 N/mm ², respectively, and with the chemical composition (in terms of element weight %): 0.43C; 0.21Si; 0.71Mn; 0.08Ni, 0.09Cr; 0.03Mo.

Friction welding was performed in a continuous drive rotary friction welding machine under the following conditions:

The specimen configuration is shown in Fig.1. The joint area in a typical specimen is shown in Fig.2 and 3.

Fatigue testing

Twenty-five specimens were tested in total. A datum S-N curve was generated first using ten specimens in the as-welded condition. For this series the applied stress ratio, R, the ratio of the minimum to maximum applied stress, was approximately 0.1.

Similarly, a group of seven specimens was used to generate an S-N curve for the stress relieved condition at R = 0.1. Thermal stress relief had been carried out on these specimens by heating them in a furnace to approximately 600°C, maintaining the temperature for one hour, and cooling in the furnace under a controlled cooling rate of less than 100°C/hr.

The remaining eight specimens were used to study the effect of flash removal, and applied stress ratio.

Flash removal was performed in a lathe using a carbide tipped cutting tool. All flash was machined off so that the diameter of the bar in the bonded area was reduced to that of the plain unwelded bar. Consequently, any axial misalignment between the two bars was amplified in these specimens, resulting in a ridge or step-like discontinuity.

The fatigue test matrix is summarised in the Table.

Results and discussion

Static strength

The results of tensile tests, using two specimens tested to failure under static conditions, suggest that with proper attention to welding parameters friction welds can match the tensile strength of plain unwelded material. It was encouraging to observe that both specimens failed in the material away from the welded area. Not surprisingly, the measured ultimate tensile strengths of 788 and 799 N/mm ², respectively, were in close agreement with that of the plain material: 804 N/mm ².

Hardness

The results of a hardness survey, performed on one of the specimens in the as-welded condition, did not show any increase in hardness arising from friction welding. Typical measurements in the bond area, the heat affected zone, and the plain material remote from the weld, were 241, 199 and 246HV5, respectively.

Residual stress

Two specimens, one in the as-welded condition and the other post-weld heat treated, were instrumented with centre-hole strain gauge rosettes and the surface residual stresses in the vicinity of the joint were measured. Of particular interest is the distribution of the component of residual stress in the axial direction ( Fig.4). These stresses, being perpendicular to the weld, are expected to have a considerable influence on the behaviour of fatigue cracks initiating at the weld.

Significantly, all axial residual stresses were compressive, with a maximum magnitude of 200 N/mm ². Elsewhere, on the surface of the parent bar away from the weld, the compressive residual stresses were smaller, but by no means negligible. Compressive residual stresses of this kind are expected to have a beneficial effect on fatigue performance, as they tend to 'clamp' any existing flaws and thus retard their growth under fluctuating applied stress.

Thermal stress relief did not eliminate the residual stresses completely, but reduced the amount of (beneficial) compressive stress substantially.

Fatigue performance

The fatigue test results are shown on a conventional S-N diagram in Fig.5. Also shown are Class B and C mean and design lines from the fatigue design code for fusion welded joints. ^[2] In all specimens with their flash intact (whether as-welded or stress-relieved) which failed in the welded area, the fatigue cracks initiated on the surface at the edge of the weld at the change in section due to the flash, see Fig.6-8.

Considering the results obtained from the as-welded specimens at R = 0.1, the mean line through the data points would be considerably shallower than the S-N curve for Class B, some specimens remaining unbroken after 10 ⁷ cycles at stress ranges up to 170 N/mm ². In fact the specimen tested at 300 N/mm ² remained unbroken in the welded region after more than 10 ⁶ cycles; the test was eventually stopped after frequent failures in the parent bar remote from the weld. The 'fatigue limit', defined as that corresponding to 10 ⁷ cycles, might well be expected to exceed 170 N/mm ² for the particular welds tested here.

The test results representing the as-welded condition were considerably better than could be reasonably expected of a similar joint produced by arc welding. For instance, if the bar had been produced by full penetration butt welding it would be assigned to Class D or E, depending on the degree of control exercised during welding; welding position; and weld process. By contrast the data from the friction welds lie above the mean Class B curve. With the shallow slope associated with the friction weld data, their superior performance compared with Class D increases as the applied stress range is reduced.

As mentioned, compressive residual stresses measured near the weld are expected to have a beneficial effect on fatigue performance. This is shown clearly by the fatigue test results. First, in the specimens representing the as-welded condition, their good fatigue performance and in particular the shallow slope of the S-N curve can be attributed largely to the compressive residual stresses at the surface of the specimen, where the fatigue cracks initiated.

Secondly, in stress relieved specimens, the reduction in compressive residual stress seems to have led to significant reductions in fatigue strength, which is in line with expectation. Indeed, the reduction in fatigue strength for these specimens was particularly noticeable at the lower stress ranges. For instance at 300 N/mm ², an as-welded specimen had remained unbroken after nearly 1.2 X 10 ⁶ cycles, whereas the stress relieved specimen failed after less than 300 000 cycles. The results suggest that Class C might be the relevant design curve for stress relieved joints in this material, but further tests are required to confirm that.

Regarding the effect of flash removal, the data so far suggest that it is beneficial, giving an improvement in fatigue strength. This is not surprising, as the discontinuity at the edge of the flash in an as-welded specimen acts as a source of stress concentration. The extent of improvement appears to increase at lower stress ranges. However, more data are required to establish whether the fatigue enhancement can be achieved consistently.

Conclusions

Examination of the fatigue behaviour of friction welded joints has shown that:

• The fatigue test results obtained from as-welded specimens at R = 0.1 exhibited little scatter, and specimens remained unbroken after 10 ⁷ cycles at stress ranges up to 170 N0mm ². This is well in excess of the corresponding value for mean Class B in BS 7608, which is 124 N/mm ².
• The mean S-N curve representing the as-welded specimens tested at R = 0.1 appeared significantly shallower than Class B. The shallow S-N curve and high fatigue limit are believed to be largely due to compressive residual stresses present at the crack initiation site.
• Stress relief by PWHT had a detrimental influence on fatigue strength. This is believed to arise from the relaxation of compressive residual stress. The results obtained to date suggest that Class C might be the appropriate design curve for this material, but further work is needed for confirmation.
• As-welded specimens tested under fully reversed loading (R = -1) had significantly higher fatigue strengths (stress range) than those tested at R = 0.1
• Limited results obtained to date suggest that flash removal increases fatigue strength. The extent of improvement increases with reduction in applied stress range.

Recommendations

Under the conditions studied in this work, friction welded joints exhibited a relatively high fatigue strength. However, this was reduced by post-weld heat treatment (PWHT). There are insufficient data to suggest design recommendations; further work is required to investigate other variables including material, welding conditions, PWHT, loading mode and flash removal.

References