How Things Fail
Data Records
| Table 1; Categories of failure examined by FSM in terms of origin. | |
| Categories | No. |
| Cranes/Towers | 25 |
| Miscellaneous | 25 |
| Shafts/Screws | 22 |
| Pressure Vessels | 19 |
| Engine Failures | 13 |
| Pumps/Pipes | 13 |
| Tanks | 12 |
| Heater/Heat Exchanger | 11 |
| Mining/Construction | 10 |
| Gears/Gearboxes | 9 |
| Bearings | 8 |
| Turbines | 8 |
| Chimney/Silos | 6 |
| Rolls | 6 |
| Wire Ropes | 6 |
| Container Equipment | 5 |
| Electric Motors | 5 |
| Refractory | 5 |
| Valves | 5 |
| Food Containers | 3 |
| Printing Machines | 3 |
| Total | 217 |
Our failure analysis laboratory has over the years received a steady stream of assignments aimed at determining the reasons why various components or machines have failed in service. These failure investigations have grown in number such that their origin and the results of our work are themselves worth of analysis. The information in this article encompasses failures we have looked at over the past six years.
![]() |
| Erosion corrosion at the entraces to copper tubing in a seawater cooler. |
Origin
This type of work is usually received via adjusters. They may call and ask us to go to an accident site and collect our own samples, or they may arrive at our laboratory at Pekeliling Business Centre with a sample in the boot of a car. The smaller failures might be a few grams in weight, the largest many tonnes.
![]() |
| This stainless steel pipe weld failed because the pre-weld fit up was poor, resulting in lack of weld penetration. |
Table 1 shows the various categories that we were able to sort the failures into. Even without considering the miscellaneous category, the diverse origin is impressive, to the extent that it is clearly difficult to provide a succinct technical answer to the question, what sort of failures do you examine? One way to describe the things that we look at would be any insured plant & machinery.
| Table 2 Failures categorised in terms of determined failure mechanism. | ||
| Failure Mechanism | No. |
% |
| Fatigue | 68 | 31 |
| Corrosion | 32 | 15 |
| Poor Design/Mfg/ Installation/Packaging |
25 | 12 |
| Operator Error | 13 | 6 |
| Unknown | 12 | 6 |
| Miscellaneous | 9 | 4 |
| Overheating | 9 | 4 |
| Wear/Deterioration | 9 | 4 |
| Brittle Failure | 7 | 3 |
| Poor Maintenance | 7 | 3 |
| Poor Welding | 7 | 3 |
| Contamination/ Foreign Objects |
5 | 2 |
| Electrical | 4 | 2 |
| Lubrication Failure | 4 | 2 |
| Pre-Existing Damage | 4 | 2 |
| Creep | 2 | 1 |
| Total | 217 | 100 |
Failure Mechanism
The failures in Table 1 have been re-categorised in terms of failure mechanism, as shown in Table 2. Fatigue and corrosion account for 46% of the failures that we are asked to investigate. This percentage is less than that reported in the technical literature*, but it must be considered that we are looking at failures that are the subject of insurance claims and chosen by adjusters, rather than random cross-section of all failures that occur in industry. Presumably many fatigue and corrosion failures are recognised as such in industry and do not become the subject of claims.
![]() |
| Fatigue failure of a pile driver ram. |





