ETL 1110-2-550
30 May 97
without-project condition. A spare parts alternative
and the type of outage, the unit will either be
should also be considered where reasonable.
repaired or rehabilitated. If the unit is repaired, then
Incremental analysis of the alternatives should be
the probability of unsatisfactory performance in
performed to allow for optimization of the number
each successive time period continues to increase.
of components to be rehabilitated.
If a unit is rehabilitated, then the probability of
unsatisfactory performance is returned to a new
c. This process permits consideration of the
condition as the equipment is considered to be
physical condition of the individual components and
restored.
the potential sequencing of repairs.
E-7. Types of Unsatisfactory Performance
d. Each simulated outage incorporates
consequences, in the form of cost resulting from
a. The analysis can include multiple types of
increased frequency of repair, increased mainte-
unsatisfactory performance with different probabili-
nance effort, and having to resort to more expensive
ties of occurrence. For example, in the Chapman
means of energy production (hydropower benefits
hydropower example, the first type could be consid-
calculations).
ered to be a catastrophic outage. For a generator
stator, this type of outage could occur if a signifi-
E-6. Incorporation of Physical and
cant number of coils failed, and a rewind was the
Economic Consequences
only possible repair. The second type of outage is
less debilitating. This outage mode consists of a
a. Several columns of the spreadsheet model
repairable coil failure.
are needed to account for the engineering reliability
b. For each type of unsatisfactory perfor-
analysis. The engineering reliability analysis
mance, outage times and costs for repair are com-
establishes the probability of unsatisfactory per-
puted. For the Chapman generators, a repairable
formance for each component for current and future
coil failure may cause an outage of 1 month at an
conditions. This probability, over time, is inserted
estimated repair cost of ,000. For a catastrophic
for each year in the modeling sequence. Current
outage, the Chapman unit is estimated to be out of
conditions and probabilities of unsatisfactory
service for a period of 24 months at a repair cost of
performance vary for each individual turbine,
||content||
,500,000.
generator, and transformer.
c. For each alternative considered, routine
b. Within each iteration, a random number is
annual O&M costs are also estimated. Under exist-
generated for each component in a given time
ing conditions, the Chapman turbine units are
period. Based on the probability of unsatisfactory
dewatered, inspected, and repaired once every
performance in that time period, the unit either
6 months. If a unit is rehabilitated, inspections are
incurs an outage or continues to operate. For
assumed to decrease in frequency with a resulting
example, if the probability of unsatisfactory perfor-
reduction in O&M costs. The time associated with
mance for turbine unit number one in the year 1993
inspections and routine maintenance must also be
is 2.19 percent, then any random number generated
accounted for in each iteration.
between 0 and 1 that is less than 0.0219 will cause
d. Subsequent columns in the spreadsheet sum
an outage to occur; any number greater than 0.0219
will indicate that the unit is still available for opera-
all unit outages for a given year. Subroutines
tion. If the unit remains operational, then the
should be incorporated in the model to prevent
probability of unsatisfactory performance in the
double counting of outage time if two interrelated
next time period increases. A random number is
components are out concurrently. If the unit is
generated for each successive time period, and the
considered to be out of service in excess of
consequences are recorded. Should a unit incur an
12 months, outage times must be carried over into
outage, depending on the alternative being modeled
the next time period.
E-3
Previous Page
