The broken pipe samples collected for the study were examined on their mode of failure based on NRC (1995) and DFID (2003) classification procedure as presented in Table 5.13. The mode of failure analysis was done in order to relate specific pipe materials to specific failure modes within the study area. Upon establishing the most predominant failure mode for a pipe material, the likely leakage exponent (N1) was estimated for that pipe material which could be used to make quick estimates of leakage flow rates. The idea of relating leakage exponent (N1)
to pipe material was also done by Farley and Trow (2003), though Greyvenstein and Zyl (2005) disproved the findings and instead indicated that leakage type is a better indicator of leakage exponent (N1). In order to confirm the findings of these researchers, the data provided in Table 5.13 was analysed in terms of pipe failure modes and results were expressed in percentages as shown in Fig. 5.6. Pictures of broken pipe samples are provided in Appendix C2.
Table 5.13 Details of broken pipe samples collected from the BWB entire water supply area for the period from November 2007 to April 2008.
Cast iron and steel did not record any breakages during the study period.
0%
10%
20%
30%
40%
50%
60%
70%
80%
Asbestos cement (AC)
HDPE Galvanized Iron (GI)
P VC Cast Iron Steel
Pipe Material
Percentage breakage for a specific pipe material
Longitudinal Circular Hole/ Pit Joint Unclassified
Fig. 5.6 Analysis of broken pipes for the data that was collected from the entire water supply area of Blantyre: November 2007 to April 2008.
Pipe material No of
breaks Longitudinal
Cracks Circular or circumferential
cracks
Hole or
pitting Joint
failures Not in any category Asbestos cement (AC) 54 7 34 9 3 1 High density
polyethylene (HDPE)
158 54 0 52 43 9
Galvanized Iron (GI) 11 0 2 8 1 0
PVC 13 6 1 0 4 2
Cast Iron - - - - - -
Steel - - - - - -
The results show that the predominant pipes broken during the study period were high density polyethylene (HDPE) pipes (67% of the total number of samples). Observations showed that most of the pipes that were broken during the study period were service connection pipes. This is confirmed by Warren (2005) who indicated that most of the times, service connections are the main source of water leakages. It should be mentioned that most of the service connections in the entire supply area were constructed using HDPE pipes, and as such this explains the reason for the high rate of breakages of HDPE pipes during the study period.
Asbestos cement (AC) pipes generally form the large part of the reticulation (GoM, 2002). The results showed that the predominant failure mode for AC pipes was circular cracking. On the other hand, HDPE pipes, which form the large part of service connection pipes, had a predominant failure mode of longitudinal cracking. Likewise, polyvinyl chloride (PVC) pipes had a longitudinal cracking as a predominant failure mode. While galvanised iron (GI) pipes’
predominant failure was found as hole (or pitting) as shown in Appendix C2. It should be mentioned that the NRRC (1995) in a study that was conducted in Canada also found that plastic pipes’ predominant failure mode was longitudinal cracking and that of metal pipes was predominantly hole (or pitting). DFID (2003) also indicated that in the case of AC pipes, soft water leaches calcite from the pipes weakening them and introducing internal stresses that lead to failure through circular cracking. Misiunas (2005) further found that generally failure modes of pipe materials would most of the times follow the same failure pattern regardless the cause of the breakage. In conclusion, the findings by NRCC and Misiunas confirm the findings of this study that every pipe material is associated with a specific failure mode.
Thornton and Lambert (2005), Lambert (2001a), Greyvenstein and Zyl (2005), and Fanner (2004b) indicated that once a failure mode is known a leakage exponent (N1) could be estimated, which could be used for quick estimates of leakage flow rates. They indicated that generally N1 values for longitudinal cracks range from 1.38 to 1.85, for circumferential (circular) crack range from 0.41 to 0.53, and for holes (or pitting) range from 0.5 to 0.524. For this reason, the pipe materials in Blantyre water supply area could be attached to expected leakage exponents. AC pipes (circumferential cracking failure mode) can therefore be expected to have N1 values in the range of 0.41 to 0.53. HDPE and PVC (plastic) pipes with longitudinal failure mode, the likely N1 values would be in the range 1.38 to 1.85, whereas GI pipes (hole failure mode), the likely leakage exponent would be around 0.5.
The findings of this study are a confirmation of findings by Farley and Trow (2003) who generalised that for plastic pipes the likely N1 values would be around 1.5, but is contrary to Greyvenstein and Zyl’s findings of 2005, which found that plastic pipes predominant failure mode is hole (or pitting). Lambert (2001) pointed out that in studies that were undertaken in UK, it was found that for the plastic pipes, the N1 value was typically around 1.5, while for GI pipes N1 was typically around 0.5. In this regard, the perception of Farley and Trow (2003) who indicated that N1 values depend on pipe material is confirmed by this study and that each pipe material has its own predominant type of failure mode.
The results of the failure modes could also be compared with the N1 values calculated in this study. The values of N1 found while investigating the relationship of pressure and leakage in
Chinyonga and BCA DMAs were 1.6 and 1.0 respectively. Chinyonga DMA’s reticulation is mainly composed of AC pipe mains and HDPE pipe service connections whilst BCA DMA is mainly composed of GI and HDPE pipe mains and most of the service connections were done using GI pipes as presented in Appendix B7. The composition of the pipe materials in the specific study areas confirms that the leakage exponents (N1) found are related to pipe materials. In summary, it should be mentioned that quick estimates of leakage exponents (N1) could be made in the Blantyre water supply area and rapid estimates of flow rates and likely impacts on the overall water balance can be made.