Veda2.0 Released!


Assigning Ramping Attribute for timeslices
#1
Hello everyone, 
I am trying to assign ramping-rate attribute for each timeslice and ran into some trouble.
My timeslices are defined like the attached image ("Fig1")
   

I have assigned ramping rate attribute ACT_UPS for each time slice like the attached image ("Fig2"),
so that, for example, for technology “ELC11PP-ANS1E”, the ramping rate (ramp-up and ramp-down) for timeslice SKN, SKD, SKP be 0.5, 0.3, 0.7 respectively.
   

From what I have understood from the formulation described in “Dispatching and unit commitment features in TIMES“ document,

ACT_UPS (UP) defined for timeslice s describes a ramp-up rate between timeslice s-1 and s.
Then is it that, for my model, if s=SKN then s-1=SKP, and if s=SKD then s-1=SKN?
If not, how can I set the order of each timeslice? (Please be generous and explain how it works in case I’ve got it wrong)

Also, in order to use ACT_UPS attribute, is it necessary to declare ACT_MINLD?
If that is not the case, what is a default value of ACT_MINLD?
I observed some significant load changes even though I have assigned ACT_UPS=0 across all time slices while doing some test-runs to see if I was on the right track.
Reasoning from the formulation, it looks like ACT_MINLD has some value though I have not declared it.

Please share your thoughts with me, and please point out things that I have misunderstood.
Thank you so much. Smile
Reply
#2
I could distinguish the following 4 questions, for which please find below my quick answers:

Q1: "Is it that, for my model, if s=SKN then s-1=SKP, and if s=SKD then s-1=SKN?"
AL> Yes, it is. The timeslices form an ordered cycle under each parent timeslice.

Q2: "In order to use ACT_UPS attribute, is it necessary to declare ACT_MINLD?"
AL> No.  Please see the documentation, Part II, Table 13, entry ACT_UPS (on page 47), and Section 6.3.6 Equation: EQ_ACTRAMP.

Q3: "What is a default value of ACT_MINLD?"
AL> None.  See Part II, Table 13, entry ACT_MINLD (on page 46–47).

Q4: "Please point out things that I have misunderstood."
AL> ACT_UPS needs also the bound type, as you can see from the documentation Part II, Table 13, entry ACT_UPS (on page 47), and Section 6.3.6 Equation: EQ_ACTRAMP. In your screenshot, there is no bound type specified.  I guess the default is UP in VEDA-FE, and so your parameters would then be for ramping up. Defining a zero value for ACT_UPS(UP) would prevent any ramping up, and so I think it is well expected that you observe some significant load changes if you assign ACT_UPS(UP)=0 across all time slices.
Reply
#3
Dear Antti,

Thank you so much for your kind reply. Smile
Your explanations were clear. I have revised the model and tested few things.
It works well and as expected!

A minor question remains:
I have read the document and it tells me that ACT_UPS are ramp rates "per hour".
Then does VEDA calculate these "per hour"s according to the "G_YRFR"?

Thanks again.
Jaewon.
Reply
#4
(14-10-2019, 12:48 PM)j1choi Wrote: I have read the document and it tells me that ACT_UPS are ramp rates "per hour". Then does VEDA calculate these "per hour"s according to the "G_YRFR"?
No, VEDA doesn't do any calculations like that.

But TIMES generates the ramping constraints according to ramping per hour, because in TIMES the idea is that the meaning of input parameters should not change when the time resolution of the model is changed.  The number of hours in each timeslice is obtained from the expression G_YRFR(r,s)*8760/Cycles(r,s), where Cycles(r,s) is the number of cycles under the parent timeslice of s (e.g. the number of days in a season). The time between successive timeslices is taken as the distance between the mid-points.
Reply
#5
(14-10-2019, 02:19 PM)Antti-L Wrote: No, VEDA doesn't do any calculations like that.

But TIMES generates the ramping constraints according to ramping per hour, because in TIMES the idea is that the meaning of input parameters should not change when the time resolution of the model is changed.  The number of hours in each timeslice is obtained from the expression G_YRFR(r,s)*8760/Cycles(r,s), where Cycles(r,s) is the number of cycles under the parent timeslice of s (e.g. the number of days in a season). The time between successive timeslices is taken as the distance between the mid-points.

I see.
Thank you for your clear reply.
It really helped me construct my model.
Thanks a million!

Jaewon.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)