We have some discussions related to the Cap2Act factor of our model with 52 weekly time-slices. The capacity unit is MW and our activity unit is in GWh.
Personally, I think that this factor is 8.76 because the relationship between MW and GWh is 8.76 since it is 8760 hours in a year.
Because a year is longer than 52 weeks, this gives a weekly activity in our model that exceeds the «real» weekly activity. To correct for this, it is proposed to adjust the Cap2Act factor to 8.736 (52*24*7).
What is the influence of using a Cap2Act factor that is lower than 8.76? And according to your opinion, what is the correct Cap2Act factor for our model?

TIMES is assumed to operate with full years, each milestone year being a representative year of its period.
Maybe look at an example: If you have an existing 1000 MW plant in the system, and we assume 8760 hours in a year, the plant would be able to produce 8760 GWh in a year at 100% capacity factor. But assume that it produced 7008 GWh in 2015, and you would like to calibrate your model according to that.
In TIMES, any process can produce at most the amount of VAR_CAP*PRC_CAPACT*NCAP_AF(A) in a year, where NCAP_AF(A) stands for the NCAP_AF, or the NCAP_AFA value if the latter is specified.
If you use 8.76 as the value of PRC_CAPACT, then you should set NCAP_AFA=0.800 in order to calibrate your model in such a way that the plant can produce at most 7008 GWh in the year 2015 (1000*8.76*0.8 = 7008).
But if you use PRC_CAPACT=8.736, then you should set NCAP_AFA=0.8022 for the same max. annual production.
Going further, if you would use PRC_CAPACT=4.38, then you should set NCAP_AFA=1.6 to calibrate your model. But for that to work, you would also have to increase your NCAP_AF values up to the value of 2 (because the default=1). That might already look rather strange: The plant availability factors should be set to 200%, because otherwise the model would not be able to reproduce the statistical production with the statistical capacity.
Consequently, I think you can use any value for PRC_CAPACT, if you just adjust the process availability factors accordingly. However, if you would like to use statistical (or real) capacity and availability factors in your model, I think you should also use the real annual hours as the basis of PRC_CAPACT, i.e. 8760 for a normal year.

Thank you for the clarification and valuable insights.
I personally think life is much easier with using 8.76, and I hope that I can convince the other modelers on this issue based on your comments.
Pernille
Antti-L
[quote pid='403' dateline='1510318853']
TIMES is assumed to operate with full years, each milestone year being a representative year of its period.
Maybe look at an example: If you have an existing 1000 MW plant in the system, and we assume 8760 hours in a year, the plant would be able to produce 8760 GWh in a year at 100% capacity factor. But assume that it produced 7008 GWh in 2015, and you would like to calibrate your model according to that.
In TIMES, any process can produce at most the amount of VAR_CAP*PRC_CAPACT*NCAP_AF(A) in a year, where NCAP_AF(A) stands for the NCAP_AF, or the NCAP_AFA value if the latter is specified.
If you use 8.76 as the value of PRC_CAPACT, then you should set NCAP_AFA=0.800 in order to calibrate your model in such a way that the plant can produce at most 7008 GWh in the year 2015 (1000*8.76*0.8 = 7008).
But if you use PRC_CAPACT=8.736, then you should set NCAP_AFA=0.8022 for the same max. annual production.
Going further, if you would use PRC_CAPACT=4.38, then you should set NCAP_AFA=1.6 to calibrate your model. But for that to work, you would also have to increase your NCAP_AF values up to the value of 2 (because the default=1). That might already look rather strange: The plant availability factors should be set to 200%, because otherwise the model would not be able to reproduce the statistical production with the statistical capacity.
Consequently, I think you can use any value for PRC_CAPACT, if you just adjust the process availability factors accordingly. However, if you would like to use statistical (or real) capacity and availability factors in your model, I think you should also use the real annual hours as the basis of PRC_CAPACT, i.e. 8760 for a normal year.
[/quote]