30 May 97
g. For each alternative, the model must be able
with any number in the range having an equal
likelihood of occurrence. Each random value is
to incorporate routine and nonroutine O&M costs
input into the spreadsheet, and the spreadsheet is
for each component over the period of analysis.
recalculated to arrive at an associated outcome.
h. The model must be able to account for
Each random trial or iteration of the spreadsheet
represents an independent "what-if" game. By
changes in generating unit efficiencies with various
generating hundreds, or in some cases, thousands of
"what-if" games, Monte Carlo sampling will gener-
i. The model must be able to incorporate the
ate the input distribution and the entire range of
consequences of events and repair/rehabilitation
scenarios in terms of changes in hydropower system
benefits and alternative construction costs. Each
E-4. Model Requirements
alternative produces different hydropower outputs,
system benefits, and O&M costs.
Basic functional requirements are established for
j. The model must be able to accommodate
the model. These requirements allow for flexibility
in the analysis, incorporation of basic assumptions,
other economic calculations such as present valua-
and the ability to change parameters as needed.
tion and amortization of costs and incorporation of
Some of these requirements are described below.
interest during construction.
a. The model must accurately reflect the
E-5. Model Operating Characteristics
without-project condition. The without-project
condition establishes a base condition from which
a. For each alternative considered, the
all other alternatives are to be evaluated.
spreadsheet is modified to simulate the specific
b. The model must be flexible enough to
engineering, operational, and economic conse-
evaluate a full range of alternatives. Alternatives
quences relative to the alternative. Monte Carlo
considered in the analysis often include: enhanced
simulation techniques are incorporated into the
maintenance, use of spare parts, a full array of
spreadsheet. This approach uses random number
rehabilitation scenarios, and, subsequently, appro-
generation to compute an expected result given a
priate timing of any rehabilitation strategy.
combination of probabilities and events. The
program sums the results of multiple iterations of
c. The model must distinguish between indi-
the simulation and produces expected values and
vidual operating components, and economic conse-
variance. Each simulation should include a mini-
quences of various alternatives, and the timing of
mum of 300 iterations. Up to 5,000 iterations may
need to be computed in some simulations.
d. The model must be able to incorporate
b. Separate simulations are conducted for the
incremental analysis of each unit and its separable
without-project and for each alternative considered
in the analysis. Simulations for the Chapman
Powerhouse example (Appendix C) should include:
e. The model must account for a project life
rehabilitation of one to four turbines; rehabilitation
(35 years is recommended) and for near-term events
of one to four generators; rehabilitation of one or
that could impact future rehabilitation strategies.
of these alternatives. The appropriate timing for
f. The model must be able to incorporate the
rehabilitation should also be evaluated. Another
alternative that should be considered is one that
analysis for each time period and each functional
uses an enhanced maintenance strategy. In many
component under evaluation.
cases this may already be implemented in the