DieselShip

Fatigue

Most engineering component failure is due to fatigue. Fatigue may be defined as the formation and gradual propagation of a crack through a material under conditions of varying tensile stress. The stress must be tensile, cracks will not open up during compressive stress, and it must vary with time, fatigue cracks do not propagate under static tensile stress. A fatigue failed component exhibits two distinct types of surface on its fractured faces. There is the shell line marks of the fatigue crack growth, this may actually be polished due to the rubbing of the crack faces together, and the very rough section of final failure where the remaining material was unable to withstand the applied load causing final failure in a brittle manner.

The relative areas of fatigue crack growth and final failure indicate the type of loading and stress application involved. A large fatigue crack indicates low stress application involved. A large fatigue crack indicates low stress high cycles type of fatigue because the crack has taken a long time to propagate through the material, the relatively small final fracture surface indicates that low stress has been applied. (an example of this is a slightly misaligned motor coupling.) A small fatigue crack surface area but large final brittle failed area indicates low cycle high stress fatigue such as may be found in a pressure vessel like a starting air reservoir.

There is a very close connection between fatigue stress and the number of cycles to failure and this may be seen on the graph for steel.

There is a very close connection between fatigue stress and the number of cycles to failure and this may be seen on the graph for steel.

From this graph it can be seen that on the sloping section there is a set number of cycles to failure for any given stress. This means that the designer must make the dimensions of his component such that the stress induced will not exceed the value for that number of cycles. Alternately he may already have a set working stress and may then determine a safe working life for his component.

Operating speed of the engine must be taken into account when determining the expected lifetime in hours, the faster the machine operates the fewer the number of hours will be needed to reach the number of cycles. This is why bottom end bolts for medium speed engines require changing after a set number of hours whilst those for slow speed engines last much longer. The idea is to replace the bolts before there is any risk of fatigue failure.

There is a value of stress, known as the fatigue limit, where the graph is horizontal. If the stress is kept below this value fatigue failure should, theoretically, never occur. In practice problems can result due to stress raisers as these act to increase the stress locally to much higher values and so cause fatigue failure to take place before it was expected or when when it was not expected at all. These stress raisers include internal material defects such as slag, gas porosity, and existing cracks, but they may also be due to sharp changes in section caused by poor design or surface damage. Fillet radii should be used to avoid such stress raisers at changes in section and often these fillets are rolled because that avoids cutting the material grains and at the same time induces a compressive stress locally which reduces the total tensile stress in the material when working stress is applied. Undercutting the bolt diameter below the root of screw threads avoids the stress raiser effect of these threads and means that the nominal bolt diameter should not be subject to stress raiser effects.

Surface damage during overhaul causes stress raiser effects whilst overtightening of bolts induces a higher static stress so that when working stress is applied very high tensile stresses are obtained causing increased risk of fatigue.

The fact that a crack is not very large on the surface of a component does not mean that serious cracking does not exist. Only by seeing the full extent of a crack internally can the situation be fully assessed, this requires the use of ultrasonic methods. Even then some idea as to the type of loading must be obtained so that the real situation may be known. A small crack may not be near to causing failure with low stress fatigue but a crack of the same dimension may be very close to causing failure with high stress fatigue. In some situations such as piston crown it may be very difficult to detect cracks due to scale and other deposits.

Welds are particularly susceptible to fatigue cracking not because they are highly loaded but because slag and other defects they contain act as stress raisers. The presence of residual tensile stress in a weld also increases the total value of tensile stress when load is applied. Vibration imparts a varying load and many fatigue failures in welds are caused by vibration. Stress due to vibration may be very low but when stress raiser effects and residual stresses are added then the effect is appreciable. Any additional loading on a weld is liable to cause problems. This was the case with the Sulzer engine when jacking screws where first employed for holding the bearing caps as these jacking screws acted upwards on the cross section of the A-frames inducing a stress which was not there previously. Fatigue cracks developed in the weld of the A-frame but this was overcome by extending the cross plates and weld. Then machining a final radius.