IFMASS10

135

Against this method of “safe life” there is the “fail safe” method. This is the case when

designing the structure in which one component may fail without causing the failure of

the structure, because the other components are capable to maintain structural integrity of

the system when the new load (stress) for the failed condition is applied at least up to the

next inspection. This is the case like bolt connection with several bolts. Similar concept

refers to the combat aircraft that have two jet engines, so that in case of one failed down

or being stroked the aircraft can fly back to the basis with the power of only one engine.

It can be seen from these examples that the practical application of the method is

based more on technical disciplines and management, that unavoidable involves the risk.

Reduction of the risk requires good knowledge of all processes based on technical

disciplines significant for the case. Decisions in this respect are always made based on

fundamental concepts and considering total cost. The latter is hard to calculate, when not

only the cost of exploitation but also of unexpected and fateful accidents are considered.

On the other hand, equipment owners today all over the world, especially in energy

branch, require assessment based on „fitness-for-purpose“ and corresponding confident

safety in application, what is reasonably based on solid experiences in this respect.

When the available inspection methods are not sufficient, other methods can be used,

which are also based on fracture mechanics. For pressure equipment is usual to perform

the test by overpressure before the service, to prove the structural integrity under the

pressure and temperature in operation. Praxis of this proof test with some overpressure

before involve the equipment in service is recognised as „good design praxis“. Behind

this stands the fact, that the endurance in the proof test conditions provides increased

safety that the component can be reliably exposed to lower loading level in service for

long time. On the other hand, the improved reliability of the structure was exhibited,

because during proof test components, which are not adequate, can be removed from

operation, without affecting the reliability of the components which survive the test.

However, formal application of this test do not attain full profit, because, in spite of all

advantages quantitative data regarding important parameters are missing, as presence and

size of cracks and discontinuities in component, lifetime of safe exploitation without

failure risk after some period, actually required level of loading in test and the real risk of

failure during the test itself. Direct benefit of this test, not limited to pressure components

and pressure loading, can be significantly increased applying the fracture mechanics in

preparation and execution of test optimisation and after that, since successful test enables

to determine safe operational conditions for the definite period of next operation.

The principle of fracture mechanics application for proof pressure testing is shown in

Fig. 12 for the case of brittle materials and single loading in test. Here is

maximal

crack dimension with which the loading

can be applied in test without failure. The

reason of proof test suppose the worst case scenario, and this means that the crack of size

is certainly present in tested component and it is taken as initial size for the evaluation

of safe life in operation after this test. For

evaluation upper values of material

toughness has to be taken giving conservative (down) values in life calculation.

Although the knowledge about the existence and distribution of remaining cracks is

limited, after successful test it is supposed that no one crack exist larger than

in the

critical location and in the weak material. Guaranteed life is than evaluated based on

calculated crack growth under service conditions based on expected mechanisms (fatigue,

corrosion, creep) up to the maximal allowed crack size

calculated for maximum

expected service loading

. It is to note, that in proof test the failure must not mean