536
2.2 Steps of the model
A mining operation requires coordinated movements of different machineries with an intention of fulfilling any of drilling, blasting, loading, and hauling activities. Deterioration of components/subsystems in an equipment may change depending on the operation type where it is
involved. Therefore, characterization of a mining equipment as a whole is not practical. A system
needs to be decomposed into its subsystem; and failure records should be clustered regarding
the subsystems. A subsystem is a collection of multiple components that gathers to perform a
specific task in a system. Therefore, any exposure to failure and the related failure modes may
be explained more explicitly following a failure and/or maintenance data classification based on
the subsystems. The current model uses random variables to evaluate functional conditions of
subsystems anywhere in the timeframe by using a probabilistic approach. For instance, in a specific simulation time, one subsystem may be detected as failed where another subsystem may be
in a deterioration, i.e. wear-out, period without any complete failure. In such a case, failure of a
subsystem may create an opportunistic time span to maintain the deteriorated but not failed subsystems. If there is not such an opportunity, the same deteriorated subsystem may be diagnosed
during regular inspections. In this basis, the following model steps evaluate the functional abilities
of subsystems, and give the resultant maintenance decisions by using discrete event simulation:
i. Generating random numbers: They are generated from the probability density functions
of surviving, i.e. time between failures (TBF), and repairing, i.e. time to repair (TTR),
durations. These numeric values provide a stochastic uptime and downtime characterization in compliance with the natural operating behavior of system.
ii. Ordering subsystems: For each simulation time, subsystems are evaluated and ordered
considering their failure conditions since opportunistic maintenance can only take place
once another subsystem fails. Therefore, subsystems with a near-failure condition is
located at the top of list with a priority. This module also ensures the completion of
simulation cycle for all subsystems for the active simulation time to provide a holistic and
complete view on the system.
iii. Checking deterministic thresholds: Three main types of deterministic thresholds are
used: Target simulation time, inspection points, and administrative breaks. In case that
any of these values is arrived at the active simulation time, the related modules are activated. The simulation is ended if the active simulation time is equal to target simulation
time. In any inspection, the subsystems with a probability of being detected as deteriorated during the inspection duration are revealed, and these subsystems are preventively
maintained. In case of any preventive maintenance, a new random TBF is generated for
the component. Administration breaks are the scheduled breaks where the production is
interrupted due to shift changes or any other administrative decision.
iv. Evaluating subsystem failure period: If the active simulation time does not give any sign
about the deterministic thresholds discussed above and if a failure is detected during anytime in the operation, this period is evaluated not just for the failed component but also for
the other subsystems in an operating condition. Because, the production interruption caused
by a subsystem failure may create an opportunistic time span for the proactive maintenance
of other subsystems. Randomly generated TTR value of the failed subsystem is compared
with the random TTR values of other subsystems. If any subsystem is in a wear-out period
and if it is decided that proactive maintenance for that subsystem can be performed within
the assigned corrective maintenance duration, then an opportunistic maintenance is applied
preventively. Any subsystem exposed to either of preventive (proactive) and corrective maintenance gets a new random TBF right after the maintenance activity.
v. Estimating production amount: System production is interrupted in a) regular inspections, b) subsystem failures that require corrective maintenance during normal operation
time, and c) administrative breaks. Other than these conditions, system is assumed to
have a production with a specific cycle time. Length of the cycle time and production
amount are introduced in compliance with the nature of production. Discrete production
amounts are gathered cumulatively to find out the total production throughout the target
simulation time.
2.2 Steps of the model
A mining operation requires coordinated movements of different machineries with an intention of fulfilling any of drilling, blasting, loading, and hauling activities. Deterioration of components/subsystems in an equipment may change depending on the operation type where it is
involved. Therefore, characterization of a mining equipment as a whole is not practical. A system
needs to be decomposed into its subsystem; and failure records should be clustered regarding
the subsystems. A subsystem is a collection of multiple components that gathers to perform a
specific task in a system. Therefore, any exposure to failure and the related failure modes may
be explained more explicitly following a failure and/or maintenance data classification based on
the subsystems. The current model uses random variables to evaluate functional conditions of
subsystems anywhere in the timeframe by using a probabilistic approach. For instance, in a specific simulation time, one subsystem may be detected as failed where another subsystem may be
in a deterioration, i.e. wear-out, period without any complete failure. In such a case, failure of a
subsystem may create an opportunistic time span to maintain the deteriorated but not failed subsystems. If there is not such an opportunity, the same deteriorated subsystem may be diagnosed
during regular inspections. In this basis, the following model steps evaluate the functional abilities
of subsystems, and give the resultant maintenance decisions by using discrete event simulation:
i. Generating random numbers: They are generated from the probability density functions
of surviving, i.e. time between failures (TBF), and repairing, i.e. time to repair (TTR),
durations. These numeric values provide a stochastic uptime and downtime characterization in compliance with the natural operating behavior of system.
ii. Ordering subsystems: For each simulation time, subsystems are evaluated and ordered
considering their failure conditions since opportunistic maintenance can only take place
once another subsystem fails. Therefore, subsystems with a near-failure condition is
located at the top of list with a priority. This module also ensures the completion of
simulation cycle for all subsystems for the active simulation time to provide a holistic and
complete view on the system.
iii. Checking deterministic thresholds: Three main types of deterministic thresholds are
used: Target simulation time, inspection points, and administrative breaks. In case that
any of these values is arrived at the active simulation time, the related modules are activated. The simulation is ended if the active simulation time is equal to target simulation
time. In any inspection, the subsystems with a probability of being detected as deteriorated during the inspection duration are revealed, and these subsystems are preventively
maintained. In case of any preventive maintenance, a new random TBF is generated for
the component. Administration breaks are the scheduled breaks where the production is
interrupted due to shift changes or any other administrative decision.
iv. Evaluating subsystem failure period: If the active simulation time does not give any sign
about the deterministic thresholds discussed above and if a failure is detected during anytime in the operation, this period is evaluated not just for the failed component but also for
the other subsystems in an operating condition. Because, the production interruption caused
by a subsystem failure may create an opportunistic time span for the proactive maintenance
of other subsystems. Randomly generated TTR value of the failed subsystem is compared
with the random TTR values of other subsystems. If any subsystem is in a wear-out period
and if it is decided that proactive maintenance for that subsystem can be performed within
the assigned corrective maintenance duration, then an opportunistic maintenance is applied
preventively. Any subsystem exposed to either of preventive (proactive) and corrective maintenance gets a new random TBF right after the maintenance activity.
v. Estimating production amount: System production is interrupted in a) regular inspections, b) subsystem failures that require corrective maintenance during normal operation
time, and c) administrative breaks. Other than these conditions, system is assumed to
have a production with a specific cycle time. Length of the cycle time and production
amount are introduced in compliance with the nature of production. Discrete production
amounts are gathered cumulatively to find out the total production throughout the target
simulation time.