Transmission efficiency between regions

[fg]

In a multi-regional electricity system model, I want to model the interconnectors with a certain transmission efficiency, say 0.98
That why I put into SuppXLS/Trades/ScenTrade_TRADEPARAMETERS.xlsx this line (amongst others)

   

which is being read into the model very well:

   

However, the results show that this constraint is not working, as the VAR_Fin and VAR_Fout for two neighbouring regions are exactly the same.
E.g. here for a transmission from Hungary to Austria in one timeslice, I would expect
VAR_FIn(HU) =   12172.7823414752  MWh
VAR_Fout(AT) = 11929.3266946457  MWh = 0.98 * VAR_FIn(HU)
but I got this:

   


After failing with COM_IE, I also tried to use both EFF and CEFF attributes for these transmission processes - unfortunately with no success (both are read properly into the model, both do not show the desired effect).
What am I doing wrong here?

Thanks in advance for any help

[Antti-L]

You might want to consider having a look at the documentation, Part II.

The description for COM_IE is: "Overall efficiency applied to the total production of a commodity in the commodity balance equation (EQ(l)_COMBAL)."  I think that makes it quite clear that it does not affect the transformation over processes at all.  And, looking at the equation formulation for EQ(l)_COMBAL in the doc, you can also see in more detail how it is applied.

EFF and CEFF are VEDA parameters, which are translated into the TIMES parameter ACT_EFF, as you can see from VEDA-FE Basic Functions → Browse → TIMES view. The table on Page 119 gives you information about using ACT_EFF for IRE processes: Can not be used.  In other words, ACT_EFF can be used for standard processes only, and if defined for IRE processes it is simply ignored.

I would recommend using IRE_FLOSUM for defining efficiencies over IRE processes. But there is also the IRE_FLO parameter, which could also be used. The drawback with IRE_FLO is that it makes the activity level different on the import/export sides, which may not be desirable.

[AKanudia]

It doesn't show up in the attribute master, but if you specify 'EFF' for an IRE process, VEDA will migrate it to IRE_FLO. Did you see IRE_FLO for this process in Browse (TIMES view)?

[fg]

Dear Antti, der Amit,

thank you for your answers.

Indeed, it looks like I've faced a huge misunderstanding concerning the term 'transmission efficiency', which is the description for COM_IE in the Attributes Master in VEDA.
I have always understood 'transmission efficiency' as the ratio of receiving end power to sending end power, exactly like in the way it is defined e.g. here:
p. 230, right above section 10.3: http://www.nct-tech.edu.lk/Download/Tech...Lines..pdf
Therefore it did not even come to my mind that something absolutely different could be meant - thank you for clarifiyng this!

Therefore I tried your hints:
IRE_FLOSUM: When I set Other_Indexes  to 'IMP,Out,ELC_ConELC' and try to read the model, VEDA claims: "Mismatch in the number of elements provided and expected in Other_Indexes column for attribute IRE_FLOSUM".
This is odd to me, as I set three elements and in the Attributes Master there are three elements to provide: ImpExp,InOut,Commodity2
Looking into the TIMES Documentation part II, p. 117, Table 19, I think I provided all the elements required: (r=AllRegions, y=Baseyear, p=TU_*, c1=ELC_ConELC, s=DAYNITE, ie=IMP, c2=ELC_ConELC, io=OUT)

IRE_FLO: Seems to be cumbersome to define manually as one would need to provide everytime region2 and timeslice2, but I want to set this indepently of the region and timeslice.
Amit, you said correctly that EFF is being migrated into IRE_FLO. For testing, I set this value to 0.5 which I can see in the TIMES View, see attachment.

However, the results are the same as above, I don't see any difference in the in- and outflows in the modelling results. 

Any further ideas?

[AKanudia]

this is a bit messy... will be cleaned up soon.

multiple items in Other_Indexes col have to be separated by '~' in FI_T tables, and by ',' in TFM tables. Both separators will work in both cases in the next update.

[Antti-L]

Oh, I was not fully aware that VEDA translates EFF into IRE_FLO.  Then you should indeed be able to use EFF for IRE processes.

But your screenshot shows DAYNITE in the Timeslice dimension. That would seem erroneous, because DAYNITE is not a timeslice.  Is this a result of the VEDA translation, or was the timeslice defined by you?  If the latter, try using ANNUAL instead (it will be levelized by TIMES).

[Antti-L]

Here is a simple example of defining transmission efficiencies with IRE_FLOSUM.
As you can see, the specification is quite easy, but you must use the separator ~ for the Other_Indexes in TFM_INS tables.  I have been defining all my transmission line efficiencies with IRE_FLOSUM, and I am quite happy with it.



For example, the transmission efficiency of the interconnector X-ELC-IND-ODA is 1-0.05 = 0.95.

[fg]

Perfect guys, setting TimeSlice to ANNUAL and using EFF did the trick.
I actually don't exactly understand why - since Amit advised me to set my time slice level, PRC_TSL to DAYNITE in the SysSettings some time ago - but I just accept Antti's note that it will be levelized by TIMES...

Actually, after seeing now how Antti defined it with negative values, I think it would also be doable in my case with IRE_FLOSUM instead.

Impeccable support, thank you, you two, again!

See you maybe next week at the ETSAP Workshop / IEW Conference 2017 ?

[Raulm]

Hey all,

I tried to set the efficiency in IRE processes using EFF, which in turn should migrate to IRE_FLO, as described by Amit.

Everything worked perfectly when adding it in "trade parameters file" (~TFM_INS) and I can see it when browsing. However, when I try to update the efficiency in a new scenario (~TFM_UPD), the attribute does not migrate to IRE_FLO and appears as an ACT_EFF attribute when browsing. I declared it exactly the same way, both in trade parameters and in the scenario file. As a result, the attribute is not updated and the IRE_FLO value remains the same.

Any ideas?

thanks,

[AKanudia]

At this point, EFF is migrated to IRE_FLO for IRE processes *only when this declaration is made in a trade scenario file*.

Mauri already suggested that this be done for regular scenarios as well, but this has not been done yet.

[Raulm]

Hi Amit,


Ok, thanks for your response.

[Raulm]

I am experiencing a weird behavior at some TB processes. I have a base scenario with both stocks and efficiency for transmission (IRE) processes. The efficiency was declared using IRE_FLOSUM, as described above by Antti (I was using COM_IE previously). Also, I have a second scenario in which some of these transmissions processes have "infinite" stocks (= 999999) and there is no efficiency losses.

The IRE_FLOSUM attribute is working perfectly. However, in particular at those processes with infinite stocks, there are simultaneous activities either entering and leaving the process, at the same timeslice and region. In the figure enclosed, under a DAYNITE timeslice, there is commodity (ELC) exchange by the same IRE process (TB_ELC_HV_CO1_CO3_1) from both regions (CO1 and CO3) to both regions (CO1 and CO3), at the same time.



Any suggestion regarding what might be happening?

[Antti-L]

Simultaneous opposite imports / export flows in the same timeslice should occur only if you have no costs, losses or other constraints that would make such flows non-optimal. In such a peculiar case, the simultaneous flows in the opposite directions indeed satisfy the optimality conditions in the (degenerate) solution. However, even superconductor transmission lines have small losses, and so I think such a lossless, costless and boundless transmission line technology would be unrealistic.

I would say your IRE_FLOSUM attribute is not working, if you have defined losses with them but still get the solution you presented. In the figure you attached I am not seeing any losses, and so maybe that case is indeed a lossless, costless and boundless case?

[Raulm]

Hi Antti,

The IRE_FLOSUM attribute is working well, there are other TB processes in which the efficiency loss can be observed. 

That's it, the one shown in the figure is a TB process which didn't have losses and bounds as I was trying to evaluate how the existence or not of some TB processes would impact the optimal solution. In other words, I was trying to evaluate how the model would work, whether some TB processes did not exist. 

And you were right, the lack of these attributes were causing the processes to work simultaneously at both directions. To overcome the issue, first I've tried to set a very small efficiency (= -0.001) in the IRE_FLOSUM, but Veda kept reading the value as zero. Then I tried to set the IRE_FLOSUM efficiency value to "-0.01", but it appeared as "-0.06" when browsing. When I set "-0.02", it appeared as "-0.07", when browsing. I don't know why it was happened, but I should mention the value had been added on a "~TFM_UPD table" and the "original value" in "trade parameter's ~TFM_INS table" is "-0.05". 

Not having success by setting efficiency values, I set a low value for the activity costs of those processes. And as you observed, having non-zero values solved the issue and I am not having simultaneously trades anymore.

thanks,

[Antti-L]

[...] first I've tried to set a very small efficiency (= -0.001) in the IRE_FLOSUM, but Veda kept reading the value as zero. Then I tried to set the IRE_FLOSUM efficiency value to "-0.01", but it appeared as "-0.06" when browsing. When I set "-0.02", it appeared as "-0.07", when browsing. I don't know why it was happened, but I should mention the value had been added on a "~TFM_UPD table" and the "original value" in "trade parameter's ~TFM_INS table" is "-0.05". 

Sounds bad.    Is this ~TFM_UPD problem happening with the recent updates of VEDA-FE? If so, maybe it is a problem with the new version? I hope the KanOrs staff can shed some light on that...