Dear @Lennart 

I am curious to understand how and why your unit is €2010/MWh. In my experience, I would normally use €/MWh as the unit of a parameter and not include a value. Or, if needed, I could add k€ or M€ or MM€ for thousands and millions of Euros.

To me, using the initial data with the 3000 should work correctly without the unit included.

Can you share you definitions of the units and maybe more on what your objective is?

Thanks!

Hi luispinto!

We use the unit €2010 and others to correct for inflation. We simulate a wide range of years, e.g. from 2010 to 2040, and use data from different sources that use different years for their prices (e.g. some input that spans multiple years is in €2010 while other is in €2015). These are then all converted to €2010 (inflation adjustment) for consistency across the model.

These are the definitions of the currency and energy quantities. EU00_EU15 etc. are parameters that define the conversion factors. These are read from the database (a procedure that is started manually, after starting the model). Could this be the problem?

DeclarationSection quantities {
    Quantity qCurrency {
        BaseUnit: €2015;
        Conversions: {
            €2000  ->€2015 : #->#*EU00_EU15,
            €2010  ->€2015 : #->#*EU10_EU15,
            €2011  ->€2015 : #->#*EU11_EU15,
            €2019  ->€2015 : #->#/EU15_EU19,
            
            ct€2000->€2015 : #->#*ctEU00_EU15,
            ct€2010->€2015 : #->#*ctEU10_EU15,
            ct€2015->€2015 : #->#/100,
            ct2015->€2015 : #->#/100,
            euro2015->€2015 : #->#,
            euro2010->€2015 : #->#*EU10_EU15,
            euro2020->€2015 : #->#/EU15_EU20,
            
            
            k€2015->€2015 : #->#*1e3,
            
            M€2015->€2015 : #->#*1e6,
            M€2010->€2015 : #->#*MEU10_EU15
        }
    }
    Quantity SI_Energy {
        BaseUnit: J = m^2*kg/s^2;
        Conversions: {
            kJ  -> J     :  # -> # * 1000 ,
            MJ  -> J     :  # -> # * 1.0e6,
            GJ  -> J     :  # -> # * 1.0e9,
            TJ  -> J     :  # -> # * 1.0e12,
            PJ  -> J     :  # -> # * 1.0e15,
            kWh -> J     :  # -> # * 3.6e6,
            MWh -> J     :  # -> # * 3.6e9,
            GWh -> J     :  # -> # * 3.6e12,
            TWh -> J     :  # -> # * 3.6e15
        }
        Comment: {
            "Expresses the value for the capacity to do work.
            
            !!m3gas -> J   :  # -> # * energy_gas,
            !ton_coal -> J:  # -> # * energy_coal,
            !ton_oil -> J :  # -> # * energy_oil"
        }
    }
}

Does that make things clear?

Yes, it does! I thought you were doing some kind of multiplication in the unit itself.

Are you able to share a small project that can reproduce the problem? This will help understand what is happening practically, since I was not able to reproduce on my computer.

I have managed to solve the problem. The issue was in the unit conversion that referenced a parameter, such as €2010  ->€2015 : #->#*EU10_EU15. The parameters EU10_EU15 where defined through another parameter, e.g. Conversion_euro('2000', '2015'), however this parameter did not contain data until the ‘Read Database’ procedure was manually executed.,In other words: at model start-up the data was missing. This presumably leads to the strange behavior.

I have solved the issue by defining the EU10_EU15 parameters explicitly as the conversion factor (i.e. defined as a number, instead of as a reference to another parameter).

Two comments:

1. It would be nice if AIMMS gave a warning if a unit of measurements conversion references non-existing data
2. It is a bit ridicules that if you use the ‘Initial Data’ wizard to define a value for a parameter using a unit, that AIMMS inserts code that doesn't compile (e.g. 3000 l€2010/MWh] ;)

If you still want to recreate the issue but are not able to do so with the above information, let me know. Thank you for taking the time to respond to my question!