Page 95 - impiantistica_3_2015
P. 95
in the field of prediction to build simulation models
(e.g. Markov, Petri Net etc.), but the Monte Carlo
modeling technique is the most widely used for the
simulation of complex and dynamic interactions
(MIL-HDBK-338B, 1998 [1]). A random number
generator is used to provide a probability of a
failure occurring which in turn is used to calculate
the time to the next failure. The Monte Carlo
model can be used to model any failure pattern
or repair characteristic for which data is available.
The outputs are a time-based availability curve
for the length of the simulation. The availability of,
and contribution to the overall downtime of each
branch, sub-system or individual equipment item
can be illustrated.
Fig. 4 - Bath tub curve when maintenance and operations records are not Case study:
available. Reliability can then be represented by eq. FLNG ram analysis
2 (Smith, 1997 [7]).
The FLNG project & project
R = e –λt (3) objectives
where A Full Supply Chain (FSC) RAM analysis has been
R is the reliability performed by Bureau Veritas on the concept stage
λ is the failure rate and t represents time of a FLNG project. The FLNG project consists of
an offshore floating facility able to treat and liquefy
Criticality feed gas from subsea wells, and store and offload
LNG onto LNG carriers as presented in figure 5.
Systems and equipment are categorised as either The subsea is composed of two different types of
‘”critical” or “non-critical” in terms of their impact on systems:
production. Failure of critical equipment results in • subsea facilities systems which are mainly
the loss of normal production associated with that
item. Conversely, failure of non-critical equipment composed of items directly linked to the gas
has no impact on production. For instance, a production from wells (i.e. Wellheads, X-mas
parallel configuration can be described with a term trees, Manifolds, Flow lines, Risers);
such as “2 × 100%”. • subsea control and safety systems which are
2 × 100% critical units do not cause production mainly composed by items linked to the control
losses unless both items have failed and are failed of subsea equipment and safety functions.
simultaneously (i.e. redundancy (Lamberson and
Kapur, 1977 [10])). Conversely, failure of either FLNG has been designed to produce about 3 mtpa
one of 2 × 80% critical units will cause normal of LNG. The FLNG facilities modelled include:
production to reduce to 80%. • production turret swivel assembly;
• raw gas treatment systems;
Fig. 5 - Full Supply Chain Monte Carlo simulation • gas dehydration systems;
• condensate recovery systems;
Production Availability There exists a number of numerical methods used • liquefaction of dry gas;
• utility systems (power generation, fuel gas and
Diagram (PAD)
Impiantistica Italiana - Maggio-Giugno 2015 93