On the relation
between NCAP_PASTI and STOCK:
NCAP_PASTI and PRC_RESID (alias STOCK) should basically
be considered as two alternative ways to define the existing capacity of a
process. You should thus not use both at the same time for any single process
(unless you know what you are doing). Internally
in TIMES, STOCK is implemented as a special kind of NCAP_PASTI, which always has
the vintage year B(T1)–1. Therefore, as you have noticed, if both NCAP_PASTI
and STOCK are used for the same process, they will sum up, whenever the vintage
years of NCAP_PASTI(v) are different from B(t1)–1.
How to model
capacities that have been commissioned in the past:
I would suggest choosing between defining STOCK(y),
y=2010, 2015, 2020, 2025,… or PASTI(v), v=1950, 1951, 1952, …, 2015. If your TIMES processes are very aggregated (i.e.
each of them represent a large number of existing plant units), I think that with
the STOCK approach it is easier to reach the correct evolution of the total existing
capacity in 2010, 2015, 2020, 2025,… Moreover, the PASTI approach is much more input
data intensive, if the number of plant vintages to be allocated to each process
is large.
For determining the correct values of STOCK(y),
y=2010, 2015, 2020, 2025,…, you should use your power plant database, and sum
up all the capacities still existing in 2010, 2015, 2020, 2025,…, according to
the plant lifetime information in the database, and set STOCK(y) for the corresponding
TIMES processes accordingly. For determining the correct values of PASTI(v), v=1950,
1951, …, 2015, you should again use your power plant database and sum up the
capacities of the plant units installed in those years, according to your aggregation
of plants to each TIMES process. Then set your PASTI(v) (and NCAP_TLIFE(v))
accordingly.
If your database is accurate, in both approaches the resulting
capacities for 2010 and 2015 should match well with the statistics. If they do
not match, you should find out why they do not match: Are there plants in the
database that may have been shut down prematurely or have extended their lifetime?
Do the statistics and the database measure the capacities in a different way
(e.g. gross vs. net capacity, nominal vs. max. hourly MW etc.)?
But if the differences are small, you can easily make a “quick
and dirty” adjustment by using the 2010 statistical value as such for STOCK(2010),
and by scaling the aggregate capacities for 2015, 2020, 2025,… according to the
2015
statistical value. Finally, you should
reach a consistent future STOCK trajectory
for the existing capacity, matching the 2010 and 2015 statistics. For the PASTI
approach, full match with the statistics may of course be a bit more difficult
to reach.
Anyway, I think that is mostly a question of input
data processing, not directly related to VEDA or TIMES, but to any energy
system model. Unless you expect that VEDA/TIMES would do that processing?
P.S.
In my case, private support is limited to technical problems with TIMES (if you think that some TIMES feature is not working as it should).
