Topic: Cost dataSubscribe | Previous | Next
Madeleine McPherson 7/7/2013
2503 Views Hi Taylor,
I am in the process of modelling a country's electricity system. The country's system operator publishes two costs: (1) The variable cost is the cost that each generator submits to the system operator which represents the cost that the generator is willing to sell his electricity for. The system operator stacks each of the generators according to their variable cost in ascending order. (2) The system marginal cost is then determined by the last generator (with the highest variable cost) that must be dispatched in order to meet demand in that hour. In my LEAP model, I would like to input the base year (2012) variable cost for each generator rather than the sum of the variable OM, fixed OM and capital costs. How would you best recommend doing this? My first inclination is to model the generator's variable cost as the variable OM cost and then leave the fixed OM cost and capital cost sections as $0- does this sound like it will produce accurate results to you? Secondly, I would like to use the merit order dispatch rule. I would like merit order to be based on generator's variable cost. If I input the generators variable cost in the variable OM section and leave the fixed OM and capital cost sections as zero, will the merit order be based on the generators variable cost?
Finally, is it possible to have LEAP tell me what the system marginal cost is for each hour based on the hourly demand profile that I have provided and each of the generator's variable cost? If so, how is this data displayed in LEAP.
Thanks very much,
Madeleine
Taylor Binnington 7/8/2013
2502 Views Hi Madeleine,
I don't see anything wrong with entering the base year variable costs as Variable OM, and am actually a little bit confused by your comment about entering the sum of the Variable OM, Fixed OM and Capital Costs - LEAP doesn't provide a field for such a value, and they aren't additive quantities anyways (for example, variable OM is per unit of energy, while fixed OM is per unit of capacity).
While you mention that you'd like to use the Merit Order process dispatch rule, I think that what you're actually talking about is the Running Cost dispatch rule. LEAP provides five different rules, you can read about them here:
http://www.energycommunity.org/WebHelpPro/Transformation/Process_Dispatch_Rules.htm
Anyways, Merit Order is something that requires you to explicitly specify the ordering of your processes, but Running Cost will dispatch them in order of increasing running cost, which is equal to Variable OM + fuel costs (after making adjustments for efficiency). Since the Running Cost rule doesn't ever look at the capital or fixed OM costs, it doesn't matter what you enter there: the dispatch order will be unaffected. You can read a bit more here:
http://www.energycommunity.org/WebHelpPro/Transformation/Dispatching_Processes_on_a_Load_Curve.htm
To answer your final question about displaying the system marginal cost, I think that the simplest way to do this is just to look at the Transformation:Outputs results view. Since you already enforced the Running Cost dispatch algorithm, all you'd have to do is look at which processes are dispatched in each year, and choose the one with the highest running cost. Since you're using this value to indicate the wholesale price of electricity in a given period, that should be all you need. You'll find that LEAP is not designed to provide an hour-by-hour breakdown of process dispatch or energy price, so if you're looking for an hourly output or price breakdown, it may not be the tool for you (perhaps investigate the HOMER model? You can also poke around here http://www.energycommunity.org/default.asp?action=71). LEAP is very good, however, at examining longer time-scale trends using scenarios.
Hope that clears a few things up,
Taylor
Madeleine McPherson 7/8/2013
2500 Views Hi Taylor,
Thanks very much for your response.
I didn't mean to imply that Variable OM, Fixed OM and Capital Costs were directly additive, but just that the total cost of a given generator would need to include these three costs.
I am having a problem with the Results I am getting. It looks like my Demand and Capacity are showing up properly, but the Transformation:Inputs, Outputs and Actual Availability don't look correct. 2012 is my base year, and Inputs and Outputs has 115.8 GWh in 2012 (which is the amount I inserted in the Historical Production tab), however then Inputs and Outputs drops to 0 for years 2013 to 2030. Also, the Transformation: Actual Availability looks correct for 2012 (all values between 0 and 1 depending on the generator and how much it ran in 2012), but then the availability drops to zero for years 2013 to 2030.
My first suspicion is that I have done something incorrect when inputting the System Energy Load Shape. For the Time Slices I am using 365 days per year (I have the hourly data, but I found that it made the model really slow), and for the Energy Load Shape, I am using % of Annual Load. Each of the days has between 0.0022 and 0.0038 in the Ave. Value column of the Yearly Shapes dialogue box. And the total Ave. Value column adds to 1.0. Does this sound correct to you? If so, can you think of why the Transformation: Inputs, Outputs, and Actual Availability all drop to zero in years 2013 (the First Simulation Year) to 2030?
Thanks very much,
Madeleine
Madeleine McPherson 7/9/2013
2490 Views Just to clarify my last post a little, I have made a couple more observations. I had been using the Dispatch Rule In Order of Running Cost. I had included Variable OM costs for each of the generators based on their 2012 bid prices, but zero for their capital cost and Fixed OM costs. This resulted in the situation described above where all electricity from 2013 (first simulation year) onward are imported instead of produced domestically. I checked the Base Year Reserves and Yield, and both are set to 1 million trillion GWh, so there should be ample reserves to generate the electricity domestically. However, when I switched the dispatch rule to Full Capacity, the problem seemed to be fixed. There were no imports to the module (which there shouldn't be as I have lots of capacity and lots of reserves), and all electricity demand was met with domestic production. Of course the exports were extremely high since all the plants were operating at full capacity, which is not representative of reality. When I changed the dispatch rule to Merit Order (and assigned Merit Orders to each of the generators), then the situation reverted to the situation when I had the dispatch rule In Order of Running Costs- i.e. electricity was imported from the base year onward instead of being produced domestically.
Do you know what my problem might be? It seems that it is related to the dispatch order, but I can't say for certain. I have checked the capacity for years 2012-2030, and that looks okay, and I have checked the base year reserves and yield (for hydro), and there should be ample supply. Do you know what other factors could be impacting the fact that electricity is getting imported instead of domestically generated when the dispatch rule is set to in order of running cost?
Thanks very much,
Madeleine
Taylor Binnington 7/10/2013
3 Likes 2489 Views Hi Madeleine,
I think you're right, this sounds to me like a problem with the load curve. Both Merit Order and Running Cost dispatch rules share the requirement of having to specify a system load curve, so I suspect that's where the problem is. You mentioned that you're using daily time slices, and that the value in each row of your Energy Load Shape is somewhere between 0.0022 and 0.0038, so that the column sums to 1.0. The numbers in this column are meant to be percentages of annual energy load, not fractions of energy load (I know they should be equivalent, but LEAP reads the value as a percentage), so the column should sum to 100.0, not 1.0. I'm actually a bit surprised that this doesn't generate an error when you try to calculate Results.
Anyways, if this is the problem then your load curve is a factor of 100 smaller than it should be, so processes are dispatched (using either Merit Order or Running Cost) to fill an area under this curve which is actually very tiny, hence you may not see anything on the Transformation:Outputs display. Out of curiosity, if you display the results as a table, not a chart, do you see small but nonzero values in Transformation:Outputs?
Right, have a look, and if this isn't the issue you'll have to send me your data set. Good luck!
Taylor
Madeleine McPherson 7/11/2013
2480 Views Thanks Taylor,
I have re-entered the load curve multiplying each of the values by 100 such that the column sums to 100. In the results table, there were very small but nonzero values in the Transformation: Outputs section, so I think you are right-only very small processes were getting dispatched.
However, when I look at the Results view at Resources: Imports, the system is importing increasing amounts of electricity in each year starting with 0.24 in 2014 and increasing to 3.14 in 2030. I have very high Base Year Reserves and Yield (both 1 million trillion GWh)in the base year as well as Additions to Reserves of 1 trillion GWh. I don't understand why the system would need to import electricity if sufficient generating capacity exists and sufficient reserves exist. Does this perhaps also have something to do with the Load Shape curve, since I am using the 2012 load shape curve for all years (2012-2030) but the demand is increasing in each year? Or do you think it is something else? Ideally, I do not want any electricity to be imported.
Thanks very much,
Madeleine
Taylor Binnington 7/11/2013
2475 Views Hi,
This could be a few things. It may be that although it appears that you have sufficient capacity, your system load curve may indicate that there are some time slices where your capacity is actually insufficient. It may also be that when you look at the Results view that show you're rising imports through 2030, you're actually viewing cumulative values (in the toolbar above the chart, click 'More...' and then select 'Cumulative' values from the drop down menu).
For diagnosing problems like this, the Energy Balance view is your friend, too. Good luck identifying the source,
Taylor