The observations above are indicative of the importance of compatibility between the pipe length and spacing in the foundation system, regardless of the pipe end conditions. For instance, if short pile lengths are preferred (due to availability of the cut-offs), an open-end system with greater spacing between the pipes is effective to bring settlement down by more than 50 %. Of course, the close-end system was apparently far less affected by the spacing between pipes, as the extra soil-base contact provided additional end resistance, enhancing its ‘floating’ capability. 4.2 Settlement Behaviour: Pipe Length and Spacing Effects The settlement – vertical stress relationships for all samples from both systems can be found in Figures 8 and 9. Note that settlement was defined as the Settlement Ratio, i.e. the ratio of settlement (ΔL) per depth of the soil bed (H), while the vertical stress was obtained by simply piding the load by the surface area of the foundation platform (i.e. 100 mm x 100 mm). These plots give more insights to the settlement control mechanism of each system, particularly with regards to the influence of pipe length and spacing respectively. The close-end system fared well in both 34 mm and 51 pipe spacing arrangements by having less or equal settlements at every stress level compared to the Control sample (Figures 8a and 9a). The trend of these plots corresponds well with the discussion made earlier on the settlement rate, where neither the pipe length nor the spacing dominated the foundation’s performance in settlement control. Nevertheless the same cannot be said of the open-end system (Figures 8b and 9b). The open-end system with 34 mm pipe spacing suffered more significant settlements at the same stresses compared to the Control sample, though the 75 mm long pipes (i.e. sample L75_34, Figure 8b) did compensate for the lack of end bearing with greater frictional resistance by the longer pipes. On the other hand, the open-end 34 mm spacing system recorded the least settlement with the shortest pipes (i.e. 25 mm), while the longer pipe systems of 50 mm and 75 mm registered similar or worse subsidence with reference to the Control sample (Figure 9b). 
The above observations point towards the dependency of the open-end system on the pile length to mobilize sufficient bearing capacity. A threshold of 75 mm length pipe is necessary, for instance, to carry imposed vertical load, even at a close spacing of approximately 1.6 times the diameter of the pipe (34 mm), centre-to-centre (Figure 9a). When spacing is increased to 2.32 times of the pipe diameter (51 mm), the longer pipes seemed to puncture the soft soil and induce excessive settlement in the foundation (Figure 9b). Shorter pipes allowed limited penetration as there was only so much internal space to accommodate the pushed-in soil, where the soil plug effectively sealed off the open ends at early stage of load application. The longer pipes, and hence larger cavities, would have permitted the soil plugs to rise higher into the pipes, resulting in more foundation subsidence. An illustration of the formation of soils plugs in the open-end system is given in Figure 10.  Figure 11 shows the correlation between Settlement Ratio and pipe spacing. As the effect of only two spacing distances were examined in this study, the data points, as expected, fall on two parallel columns in the plot. The close-end system clearly displayed better settlement control, registering ΔL/H of 0.05-0.08. The open-end system endured greater settlements, as is evident in the ΔL/H values that charted higher range, i.e. 0.06-0.10, except for L25_51. This exception matches the observations and discussions made in the previous paragraph with reference to Figures 9 and 10, which was attributed to the soil plug penetration depth effect. The Young’s modulus (E) was derived by taking the settlement of the foundation systems at the final loading stage of 1.25 kPa, and plotting a linear regression line through the settlement data points and the origin. In the plot of Settlement Ratio (ΔL/H) versus E (Figure 12), it appeared that the stiffness of the foundation system is inversely related to the settlement, i.e.  ΔL/H = 0.16/E. In other words, the stiffer a foundation is, the less settlement can be expected of it. Again, data points of the curve in Figure 12 correspond well with the settlement behaviour elaborated above. “阿卡尔基础”软土施工沉降控制英文文献和中文翻译(4):http://www.751com.cn/fanyi/lunwen_54419.html