Ratio between output commodities

[Antti-L]

If the process should be at the DAYNITE level, perhaps the easiest way of defining ANNUAL shares for the outputs is to use a non-standard FLO_SHAR, with 'ACT' as the commodity. The following example illustrates the differences of defining ANNUAL level shares for a DAYNITE and ANNUAL level process:

The ANNUAL level share constraints shown above for the DAYNITE and ANNUAL process are equivalent.

The nice thing about the ANNUAL level process is that the load curves defined for the outputs will be automatically reflected in the input flows. And you can still define e.g. timeslice-specific efficiencies for the inputs, as illustrated above (CEFF for the WP timeslice). The normal COP coefficients and annual availability factors can be defined separately for heating and cooling, as also shown above.

[saleh]

Thank you Antti for your explanation. I have another question, when I use heat pump technologies in the results I see that they only use exactly 50% of their AFACs for each commodity (cooling and heating). when I change the AFA of heat pump to 2, then for each commodity their full amount of AFC is reached. I am confused, I thought that each of produced commodities can use up the entire stock of that process not half of them. Would you please explain how that works?

Thanks,

Saleh

[Antti-L]

I am a bit confused: Do you mean that the technologies produced exactly 50% of heating and 50% of cooling?  I don't see why that should happen, unless you are fixing the annual output shares to be 50% each.

Concerning commodity-specific availability factors, they can be used when the capacity could be used for producing several different commodities with flexibility in the output shares. Perhaps a good example is a car, which could be used to produce either long-distance or short distance travel.  Using an availability factor A for long-distance travel and an availability factor B for short distance travel, the maximum annual output would be CAP*A of long-distance travel, or  CAP*B of short distance travel, or any combination between these extremes.

I admit that applying the same approach directly to the heat pump technology does not work well, because the combined ANNUAL availability factor cannot be derived as a linear combination of the outputs. Instead, it would reach its maximum value at some point in the middle.  I forgot to consider that in my examples, but I think this problem applies to both the DAYNITE and ANNUAL level approach.  To improve the representation in that respect, you could use a third output, e.g. "HCMIX", which would assume a reasonable predefined ratio of heating and cooling, and would thus have a larger AFAC availability factor.

But of course, if you only assume a fixed ratio between the annual heating and cooling output, as you seem to imply in your question, there would be no need to use commodity-specific availability factors at all.  They are only useful when there is some flexibility in the outputs.

[saleh]

The problem I have is not a fixed ratio on the annual output. I have a flexible annual output ratio on the outputs (FLO_SHAR~UP=0.6 for both cooling and heating).

The issue is when I calculate the used AFAC of the outputs after running the model (which is CF) for heating and cooling, it is exactly equal to 50% of defined AFAC for each of them. For example, AFC of heating is 0.0064, the CF that I observe is exactly equal to 0.0064*0.5=0.0032 in all periods. This is different from FLO_SHAR constraint which is not exactly 0.5 and it puts constraint on the output ratio of heating and cooling. I don't have any other constraint and I can't understand why this happens. In order to prevent this to happen and force the model to use the full amount of AFAC, one way is to put AF of the process equal to 2. But I assume the model should be able to reach CAP*AFAC for each of outputs commodities, not CAP*AFAC*0.5. Is that correct?

[Antti-L]

I am sorry but I am not able to grasp the full picture from the details you are reporting.

For example, you refer to both AFC and AFAC, and you say that "AFC of heating is 0.0064". I assume that you are aware that AFC requires also a timeslice level?  But I cannot see you giving any information about the timeslice level of this AFC.

Concerning your question about the ability of producing multiple outputs without impact on the required capacity when using commodity-specific availability factors, I am afraid the answer is no. Assume that you have a flexible machine that can produce, say, apple juice, orange juice and banana juice, with a nominal maximum capacity of 1000 litres per hour. Due to differences in the peeling and processing of the fruits, and taking into account maintenance, the machine could actually produce at most 900 litres of apple juice, or 500 litres of orange juice, or 700 litres of banana juice per hour. It would never be able to produce more than 900 litres per hour.

However, it seems you are suggesting that the machine should be able to produce 2100 litres of juice per hour, by producing all three types of juice at the same time, each with the assumed maximum fruit-specific availability factor. The total output would thus be over two times the nominal capacity. But if producing only orange juice, the max. output would still reduce to only 500 litres per hour?  Is this the correct interpretation of what you expect?

The TIMES commodity-specific availability factors do not work that way. With them you can only differentiate the availability factor according to the actual mix of outputs. For example, using equal availability factors for 10 different outputs would be equivalent to just using the same value for the standard availability factor; it would not increase the maximum output ten-fold larger compared to the single output case, as you seem to suggest.

[saleh]

I would appreciate it if you can help me with this. As I said earlier, I want to model heat pump technologies that can produce both space heating and space cooling service demand. AFACs represent the annual availability factor of those service demands. This is how I define this technology. 

As you explained in juice machine example, I expect if 1 PJ of heat pump capacity installed in a year, the model should be able to use 0.5 PJ of capacity for space heating and 0.5 PJ of capacity for space cooling. Since both related AFACs are less than 0.5 (0.28 and 0.17), I expect the model to produce both of these service demands with capacity factor of 0.28 and 0.14.

when I run the model without specifying FLO_SHAR for each commodity, model will not produce both space heating and space cooling in one year. It only produces heating with its full capacity (0.28).

But, when I specify FLO_SHAR for the commodities (0.28/(0.28+0.14) and (0.14/(0.28+0.14)) the model produce both commodities with the capacity factor of (0.28/2 and 0.14/2) (It seems the model assumes there is 0.5 PJ available out of  1 annual PJ of installed capacity for each commodity). This does not make sense for me, I want the model to be able to produce both commodities at their nominal level of AFACs in the same year (since heat pumps should produce heating in winter and cooling in summer), and more importantly, WITHOUT specifying FLO_SHAR ( I want to have more than two output commodities and I am not able to calculate the exact SHAR-FLO for each of them) .

Would you please help me how I can model this?

Thanks,

Saleh

[Antti-L]

I am sorry that you still seem to be stuck with trying to use the standard NCAP_AFAC, although I have tried to explain that it is not the right approach for accomplishing what you want.

I have already explained that when using the standard NCAP_AFAC, the combined availability factor is derived as a linear combination of the outputs. I have also noted that this does not work well for technologies like heat pumps, for which the combined ANNUAL availability factor cannot be derived that way, but it would reach its maximum value at some point in the middle of the range of possible outputs mixes.  In other words, what you want is quite different from a standard NCAP_AFAC, unless you introduce a third output representing a predefined mix, which may become cumbersome. The differences are illustrated in the following graph (denoting by Saleh-AFA the combined availabilities you want, and by AFAC-AFA those that you get with standard NCAP_AFAC with the two outputs, assuming AF factors 0.28 and 0.14 for heat and cool):

As you can immediately see form the graph, if you fix the output share of heat to 2/3 (as you apparently did), you will inevitably get a combined output availability of 0.21 if you use the standard NCAP_AFAC approach, just as you reported. So, the model works exactly as you would expect.

You can also see that in order to model the availabilities in the way you want, instead of using the standard NCAP_AFACs you basically need to define two independent flow-capacity equations, as follows (assuming the AF factors were 0.28 and 0.14):

FLOW(CHPTASHP06,COMSPH,ANNUAL) ≤ 0.28 × CAP(CHPTASHP06)

FLOW(CHPTASHP06,COMSPCS,ANNUAL) ≤ 0.14 × CAP(CHPTASHP06)

It is very easy to define these two equations with the user constraint facility.

Would you agree about that?

Finally, in TIMES v3.7.0, you can also use NCAP_AFAC in a non-standard way to accomplish the same.  You just need to add a column AFAC~ACT~0 and put a value –1 under it for the process. This addition was implemented specifically to address your problem even without defining UC constraints or additional dummies.

[saleh]

Hi Antti,

Thank you for your explanation. Now it makes sense for me how TIMES works. Unfortunately, it seems I AFAC~ACT~0 does not work for me, I am using VEDA-FE v 4.4.7 and TIMES v370!

Saleh

[AKanudia]

"does not work" - does VEDA not read your specification well or are the results different from your expectations?

If your specification is read OK, please post a picture of the browse screen that shows all indexes of this attribute (for the process you are working with).

[saleh]

Actually VEDA does not read my specs.

[AKanudia]

pl post a pic of the declaration you have made in Excel.

[saleh]

[AKanudia]

do this instead:

your spec will work after the next update of VEDA-FE

[saleh]

Thank you. It imports well and the results make sense.

[vincedh]

I am sorry
that you still seem to be stuck with trying to use the standard NCAP_AFAC,
although I have tried to explain that it is not the right approach for accomplishing
what you want.

I have
already explained that when using the standard NCAP_AFAC, the combined
availability factor is derived as a linear combination of the outputs. I have
also noted that this does not work well for technologies like heat pumps, for
which the combined ANNUAL availability factor cannot be derived that way, but
it would reach its maximum value at some point in the middle of the range of possible outputs mixes.  In other words, what you want is quite different from a standard NCAP_AFAC, unless you introduce a third output
representing a predefined mix, which may become cumbersome. The differences are
illustrated in the following graph (denoting by Saleh-AFA the combined availabilities
you want, and by AFAC-AFA those that you get with standard NCAP_AFAC with the two
outputs,
assuming AF factors 0.28 and 0.14 for heat and cool):

As you can
immediately see form the graph, if you fix the output share of heat to 2/3 (as
you apparently did), you will inevitably get a combined output availability of
0.21 if you use the standard NCAP_AFAC approach, just as you reported. So, the
model works exactly as you would expect.

You can
also see that in order to model the availabilities in the way you want, instead
of using the standard NCAP_AFACs you basically need to define two independent
flow-capacity equations, as follows (assuming the AF factors were 0.28 and 0.14):

FLOW(CHPTASHP06,COMSPH,ANNUAL) ≤ 0.28 × CAP(CHPTASHP06)

FLOW(CHPTASHP06,COMSPCS,ANNUAL)
≤ 0.14 × CAP(CHPTASHP06)

It is very
easy to define these two equations with the user constraint facility.

Would you
agree about that?

Finally,
in TIMES v3.7.0, you can also use NCAP_AFAC in a non-standard way to accomplish
the same.  You just need to add a column
AFAC~ACT~0 and put a value –1 under it for the process. This addition was
implemented specifically to address your problem even without defining UC
constraints or additional dummies.

Hello Antti,

I have a question concerning the plotting of the resulting NCAP_AFAC combination with the two outputs. You said it was in principle linear, so I would have guessed the combined output availability to be (0,28*heating + 0,14*cooling) / (heating+cooling). Hence with a share of heat of 2/3, it would give an availability factor of 0,23 and not 0,21.
I must have missed something, but I don't see what. Could you explain how you obtained the combined availability factor ?

Many thanks,
Vince