CRACK PROPAGATION IN ALUMINIUM GAS CYLINDER NECK MATERIAL AT CONSTANT LOAD AND ROOM TEMPERATURE 常温恒载条件下铝制气瓶瓶颈材料的裂纹扩展
H. L. STARK and R. N. IBRAHIM H. L. STARK 和 R. N. IBRAHIMSchool of Mechanical Engineering, University of New South Wales, PO Box 1, Kensington, NSW 2033, Australia 新南威尔士大学机械工程学院，澳大利亚 悉尼市肯辛顿区，邮政信箱 1，2033
Abstract 摘要
Sustained tensile load tests were carried out at room temperature on small circumferentially cracked cylindrical specimens taken from the necks of aluminium gas cylinders. Comparison was made of the resulting time to failure of specimens of aluminium alloys 6351,6061 and 6351 with lead impurity levels exceeding 200 ppm . The effect of superimposing a small number of interruptions to the sustained load was also examined. With a few exceptions the 6061 alloy withstood the sustained loading for significantly longer periods than the 6351 alloy. The presence of lead substantially reduced the time to failure, as did also the addition of a small number of fatigue cycles. 在室温下，对来自铝制气瓶颈部的小圆筒状试样进行了持续拉伸载荷试验。比较了 6351、6061 和含有超过 200 ppm 铅杂质的 6351 铝合金试样的失效时间。还研究了叠加少量间断对持续载荷的影响。除了少数例外情况，6061 合金比 6351 合金能够承受更长时间的持续加载。铅的存在显著缩短了失效时间，同样，加入少量疲劳循环也会缩短失效时间。
INTRODUCTION
Cracks have been observed in the necks of a number of aluminium gas cylinders. These cracks are in a radial plane and not associated with the large number of pressure cycles typical of fatigue failure. The authors have earlier demonstrated[1] that small cracked specimens of the 6000 series aluminiums typically used in such applications can experience crack growth to failure when subjected to sustained unvarying load giving subcritical stress intensities at room temperature. The cracked specimens were subjected to a constant load at room temperature and the time to failure recorded. The cracked specimens used in this study were small cylindrical circumferentially notched and cracked specimens as previously described by the authors[2,3]. The size of these specimens allowed five to be machined from one gas cylinder neck (scuba sized) together with one small tensile specimen of similar size. Four specimens only, taken from one piece of extrusion, were tested in the initial study[1]; in this study some hundreds of specimens taken from approximately 80 gas cylinders have either been tested or are still under test. This particular study has, in addition to confirming the earlier pilot study[1], been concerned with determining the difference between the performance of 6351 and 6061 aluminium alloys, and determining the effect of lead impurities on this mode of failure. Of further interest has been the effect on the time to failure of a small number of fatigue cycles superimposed on the sustained load applied to these specimens. 在许多铝合金气瓶颈部观察到了裂纹。这些裂纹位于径向平面内，并未与典型的疲劳失效所伴随的大量压力循环相关。作者之前已经证明[1]，使用于此类应用的 6000 系列铝合金的小尺寸裂纹试样，在承受持续不变的载荷时，可以在室温下经历亚临界应力强度下的裂纹扩展直至失效。这些裂纹试样在室温下承受恒定载荷，记录了失效时间。本研究中使用的裂纹试样是小尺寸圆柱形环向缺口裂纹试样，如作者先前所述[2,3]。这些试样的尺寸允许从一个气瓶颈部（潜水瓶尺寸）中磨制出五个试样，同时还有一个相似尺寸的小拉伸试样。在最初的试验中[1]，仅测试了来自一个挤压件的四个试样；而在本研究中，从大约 80 个气瓶中测试了数百个试样，或者仍在测试中。 除了确认之前的先导研究[1]外，本研究还致力于确定 6351 铝合金和 6061 铝合金在这方面的性能差异，并研究铅杂质对这种失效模式的影响。此外，还对这些试样上叠加少量疲劳循环后失效时间的变化产生了兴趣。
EXPERIMENTAL 实验
From each sample of material one cylindrical tensile specimen having a gauge length of 55 mm and a diameter of 7 mm was machined such that its axis was normal to, and length symmetrical about, a radial plane in the gas cylinder neck. Typically a further five specimens of the K1cK 1 c configuration (A in Fig. 1) were taken from each sample of material; these being similarly oriented with respect to a radial plane in the gas cylinder neck. One of these five specimens was used to measure K1cK 1 c and the further four were fatigue cracked and subjected to constant load and the time to failure recorded. This sustained load testing was carried out at 20^(@)C.K1c20^{\circ} \mathrm{C} . K 1 c, and K1K 1 at the start of the sustained load test, were calculated using the method described in refs [4, 5]. 从每种材料样品中，加工出一个长度为 55 毫米、直径为 7 毫米的圆柱拉伸试样，使其轴线垂直于并沿气瓶颈部的径向平面对称。通常，从每种材料样品中还取出了五个 K1cK 1 c 配置的试样（图 1 中的 A），这些试样相对于气瓶颈部的径向平面也具有相同的取向。在这五个试样中，其中一个用于测量 K1cK 1 c ，其余的四个进行疲劳裂纹处理，并在恒定载荷下测试，记录失效时间。这种持续载荷测试在开始时进行了 20^(@)C.K1c20^{\circ} \mathrm{C} . K 1 c ， K1K 1 的计算方法如参考文献[4, 5]中所述。
Sustained load testing was carried out using 60 of the loading jigs described in Fig. 1. After cracking the specimen in a rotating beam fatigue test machine as described earlier[2], each sustained load specimen was assembled in a jig as indicated in Fig. 1. Springs GG were compressed by application of force to components E and F , thus allowing collar D to be tightened and so applying tension to the specimen on release of the compressive forces applied to components EE and FF. On re-application of load to EE and FF the resulting relationship between load and displacement had 持续加载测试使用了 60 个如图 1 所示的加载夹具。在旋转梁疲劳试验机上按照之前所述的方法[2]对试样进行开裂后，每个持续加载试样都按照图 1 所示组装在夹具中。通过施加力到部件 E 和 F，压缩弹簧 GG ，从而使夹环 D 紧固，从而在释放施加到部件 EE 和 FF 的压缩力后对试样施加拉力。重新施加力到 EE 和 FF 后，得到的载荷与位移之间的关系为
Fig. 1. A total of 60 such test rigs were used in this investigation. A: small circumferentially notched and cracked specimen. O.D. 9.5 mm , overall length 95mm,60^(@)95 \mathrm{~mm}, 60^{\circ} notch of 7 mm root diameter. B: split collets both ends. C: threaded ring with tapered bore to house collets at lower end of specimen. E: cylindrical body of test rig having a sliding fit bore over the specimen. F: mounting ring for collets on top end of specimen. G: six Belleville spring washers giving a total spring stiffness of 4120N//mm4120 \mathrm{~N} / \mathrm{mm}. 图 1. 本研究中共使用了 60 个这样的测试装置。A：小圆周缺口和裂纹试样，外径 9.5 mm， 95mm,60^(@)95 \mathrm{~mm}, 60^{\circ} 缺口根部直径 7 mm。B：分体卡环，两端开口。C：带有锥形孔的螺纹环，用于容纳试样下端的卡环。E：测试装置的圆柱体部分，其滑动配合孔套在试样上。F：安装在试样上端的卡环座。G：六个贝尔维尔弹簧垫圈，总弹簧刚度为 4120N//mm4120 \mathrm{~N} / \mathrm{mm} 。
a change in slope when the specimen become unloaded, so indicating the initial tensile load in the specimen. Approximately six such load applications were required to adjust the tensile load applied to the specimen to the desired values. The tensile loads applied to the sustained load specimens were in the range 8000 to 12,000N12,000 \mathrm{~N}. 标本卸载时斜率发生变化，表明标本初始的拉伸载荷。大约需要六次这样的载荷应用来调整施加在标本上的拉伸载荷至所需值。施加在持续载荷标本上的拉伸载荷范围为 8000 至 12,000N12,000 \mathrm{~N} 。
After setting the load on the specimen each jig was attached to an hour meter that recorded elapsed time until fracture of the specimen (see Fig. 2). 在设定标本的载荷后，每个夹具都连接到一个计时器，记录直至标本断裂的时间（见图 2）。
The yield strength of each sample of material was determined from the tensile test specimen for that sample. After each specimen had failed, its stress intensity at the start of its sustained load test was determined as previously described[4, 5]. 每种材料的屈服强度是从该材料的拉伸试验标本中确定的。在每个标本断裂后，根据先前的描述确定其持续载荷测试开始时的应力强度[4, 5]。
Specimens that did not meet the requirements for valid plane strain fracture were deleted from the results; the validity limits used were as earlier described[3] excepting that a further limit was applied. The validity limits as described in refs [3,6] required the fatigue crack depth to be at least twice the Irwin plastic zone correction factor in depth, and that the average stress across the ligament after fatigue cracking should not exceed 2.5 times the yield strength. These limits have been extended to cater for those cases where the final ligament is eccentric to the specimen centreline; the maximum nominal stress considered is then a combination of tensile and bending stress and this should not exceed 2.5 times the yield strength. (The authors are continuing work on validity limits for the eccentrically cracked specimens, but are using the following relationship as an interim measure, in the belief that the final version will closely approximate the following.) That is, to avoid general yielding 不符合有效平面应变断裂要求的试样从结果中删除；有效性的限制条件如前所述[3]，除了增加了进一步的限制条件。如参考文献[3,6]所述的有效性限制条件要求疲劳裂纹深度至少为伊文思塑性区修正因子的两倍，并且疲劳裂纹后梁平均应力不应超过屈服强度的 2.5 倍。这些限制条件已扩展以涵盖最终梁相对于试样中心线偏心的情况；在这种情况下，考虑的最大名义应力是拉应力和弯曲应力的组合，且不应超过屈服强度的 2.5 倍。（作者们正在继续研究偏心裂纹试样的有效性限制条件，但在此期间使用以下关系作为临时措施，相信最终版本将接近以下关系。）也就是说，为了避免整体屈服
2.5 xx Sty >= (4P)//(pid^(2))+(16 P epsilon)//(pid^(3)),2.5 \times S t y \geqslant(4 P) /\left(\pi d^{2}\right)+(16 P \epsilon) /\left(\pi d^{3}\right),
where 其中
Sty == yield strength Sty == 屈服强度 P=P= tensile force applied to the specimen P=P= 作用在试样上的拉力 d=d= equivalent diameter of the ligament area after fatigue cracking d=d= 疲劳裂纹后韧窝面积的等效直径 epsilon=\epsilon= distance between the ligament centroid and the specimen centreline. epsilon=\epsilon= 韧窝质心与试样中心线的距离。
Further, for such eccentric ligament cases the fatigue crack depth considered for the purposes of validity was taken to be the greatest depth of crack. 此外，在此类偏心韧窝的情况下，用于有效性的疲劳裂纹深度被取为裂纹的最大深度。
RESULTS 结果
The results relating to the 6351 alloy (taken from material having a lead impurity content typically less than 10 ppm ), are summarised in Fig. 3. Twenty-six valid initial K1cK 1 c results were 6351 铝合金（含铅杂质通常少于 10 ppm）的结果总结在图 3 中。共有 26 个有效的初始 K1cK 1 c 结果。
Fig. 2. Thirty test jigs were connected to hour meters as shown in order that the elapsed time to failure of the specimen could be recorded. A further 30 test jigs were additionally equipped with displacement transducers to monitor the growth in length of the specimens as the test progressed. 图 2。30 个试样夹具连接到计时器，以便记录试样失效所需的时间。另外 30 个试样夹具还配备了位移传感器，以监测试样在测试过程中长度的增长。
