Veda2.0 Released!


Strange behavior plant dispatch
#1
Hi,

Something bizar is happening in a testcase I'm experimenting with to test some features for a bigger model.

The test model consists of a exogenous demand for electricity, and a limited amount of generating technologies: coal fired plants, and gas turbines.

Characteristics of both technologies, as well as the IRE processes providing the respective fuels are given below:

TechName *TechDesc Comm-IN Comm-OUT CUM COST EFF STG_EFF STG_LOSS CEFF-I Stock~2010 Stock~2020 Stock~2030 Stock~2040 Stock~2050 Stock~2060 AFA FIXOM VAROM Life CAP2ACT
*Technology Name Technology Description Input Commodity Output Commodity Reserves Cumulative Value Extraction cost or Import Price or Export Price Efficiency Cyclic efficiency of storage device Annual Storage loss (as a fraction of storage content) Commodity Input Efficiency Existing Installed Capacity Existing Installed Capacity Existing Installed Capacity Existing Installed Capacity Existing Installed Capacity Existing Installed Capacity Annual Availability Factor. Fixed O&M Cost Variable O&M Cost Lifetime of Process Capacity to Activity Factor
EPLT_COAST10 COA ELC 0.36 17 17 17 17 17 17 1 80 31.536
EPLT_GASGT10 GAS ELC 0.39 2 2 2 2 2 2 1 80 31.536
DMD_ELC ELC ELCDEM 1 50 50 50 50 50 50 1 100 31.536
IMP_COA COA 4
IMP_GAS GAS 99999
EPLT_PUMPST ELC ELC 0.95 0 2 2 2 2 2 2 1

The demand for each TS is given below. I use 12 TS: 4 SEASONAL (SU, FA, WI, SP) and 3 DAYNITE TS (D, P, N)

CommName Demand SUD FAD WID SPD SUP FAP WIP SPP SUN FAN WIN SPN
*Demand Commodity Name Demand Value Load Curve by time slice
ELCDEM 315.36 0.1 0.125 0.15 0.125 0.06 0.075 0.09 0.075 0.04 0.05 0.06 0.05
*Check 1
G_YRFR 0.1 0.1 0.1 0.1 0.05 0.05 0.05 0.05 0.1 0.1 0.1 0.1
Demand per TS in Baseyear [PJ] 31.536 39.42 47.304 39.42 18.9216 23.652 28.3824 23.652 12.6144 15.768 18.9216 15.768
Average power per TS in baseyear [GW] 10 12.5 15 12.5 12 15 18 15 4 5 6 5

As the average power per TS in combination with the stock of existing plants indicates, only in the WIP TS, the demand can't be met without using the extremely expensive gas turbine and the storage device. As expected, the results show that in this TS, the pumped storage is activated to it's maximal potential (limited in flow through an external scenario file), and some help from the GT is needed. As expected, the GT is used as little as possible in this TS (see results below).

Period 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010 2010
Attribute Process\TimeSlice FAD FAN FAP SPD SPN SPP SUD SUN SUP WID WIN WIP
EQ_CombalM - 11.11111 11.11111 11.11111 11.11111 11.11111 11.11111 11.11111 11.11111 11.11111 11.11111 11.11111 22756.41
VAR_FIn DMD_ELC 39.42 15.768 23.652 39.42 15.768 23.652 31.536 12.6144 18.9216 47.304 18.9216 28.3824
VAR_FIn EPLT_PUMPST 1
VAR_FOut EPLT_COAST10 39.42 15.1412 23.652 39.42 15.1412 23.652 30.9092 12.6144 18.9216 47.304 19.9216 26.8056
VAR_FOut EPLT_GASGT10 0.6268 0.6268 0.6268 0.6268
VAR_FOut EPLT_PUMPST 0.95

However, for some unclear reason, the GT is also used in other TS, where to Coal Plant should have no problem of delivering the demand. It is remarkable that the output of the Gas turbine is identical for all TS in which it is used. The price (EQ_Combal.M) also indicates that an additional unit of demand could be delivered by the coal plant (11.11= 4/0.36). Anyone who can explain these bizar results?

Some additional remarks:
I would expect the price in the WIP to reflect the cost of producing an additional unit of electricity with the gas plant (99999/0.39), but this is not the case.
The price of using the gas plant (99999/0.39) exceeds the cost of the dummy variables, yet, these are not used. Why could this be?
Reply
#2

If you don't find explanations to your questions, I think you could post this test model here (by putting the *.DD files and the *.RUN file into a ZIP file and uploading it here; you should find these files in the GAMS_WrkTIMES folder). Then it would be easy for anyone to reproduce your results and find the explanation. (For one, I would be willing to have a look at it).

EDIT: Are you sure you don't have dummy imports of GAS? I think it would explain the peculiar results, assuming that GAS or the dummy imports are defined at the SEASON level.

Reply
#3
Thank you for the response Antti!

It seems that the strange results are indeed related to the use of dummies (when I lower the gas price, while maintaining the merit order of the involved plants, the behavior becomes normal).

This brings me to the dummies. I know that these dummies are incorporated to facilitate the search for possible causes of infeasibilities in the model. However, their implementation is nut fully clear to me.

I maintained the dummies from the DEMO model:

~ImpSettings
Option Value
Check #DIV/0 and #REF errors in Templates 1
Create Dummy Imports for Energy and Material Commodities 1
Create Dummy Imports for Demands 1
Generate Vintage Bounds 0

~TFM_UPD
TimeSlice LimType Attribute Year Other_Indexes AllRegions BE Pset_Set Pset_PN Pset_PD Pset_CI Pset_CO Cset_Set Cset_CN Cset_CD
ACTCOST 2222 IRE IMP*Z
ACTCOST 8888 IRE IMPDEMZ

When looking at the TIMES view, IMPDEMZ and IMPNRGZ are defined on an ANNUAL TSlvl.

So my questions about the dummy import processes are:

a) Where, and how exactly are these dummy processes defined? Is this something VEDA does?

b) So these dummy imports are on the ANNUAL level. Yet I don't understand why the total dummy imports would not equal the amount needed to be able to deliver demand in the peak TS (WIP). As can be seen in the original question, the Gas turbine utilizes 4 times the necessary amount, and so does IMPNRGZ. I might be getting confused here, but I would expect that the import dummy IMPNRGZ would behave similar to the other IMP processes (IMPGAS and IMPCOA, defined at SEASON TSlvl in my case)?
Reply
#4

If your GAS commodity is defined at the SEASON level (which I suspect), then the ANNUAL GAS imports from the dummy process will be evenly distributed to the four seasons, and this explains the extra gas use in the GT process. In other words, the dummy import is the cheapest option to supply GAS, but it can only supply GAS evenly throughout the year, unlike IMPGAS, but the latter is more expensive.  And because the extra GAS is available in the other seasons, the model of course takes it into use.

The dummy import stuff is something that VEDA does automatically according the options you have chosen. You should always try to get rid of any dummy imports being active in the model results, because they indicate modeling errors.


Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)