Conversions of existing technology - feedback request

[LukeSearle]

Hi, 

Many of our existing industrial coal boilers are being upgraded to work with biomass instead. It is obviously cheaper to upgrade existing technology rather than invest in new biomass boilers, where possible. I would like to allow for this interaction in our model, but I am not sure how to best represent the conversion cost of existing processes. 

My current idea is to create two dummy processes to represent the fuel inputs for the existing boilers. One could accept coal, the other biomass, and each would output a dummy commodity that could only fuel the existing boilers. Then I could ensure the dummy coal process capacity matched the existing boilers (with no cost), and the dummy biomass process could be invested in with NCAP_COST, which would represent the conversion cost of existing coal boilers. I would additionally add a third process, which did not take dummy inputs, and simply represents a new biomass/coal boiler, which has a different cost profile but can accept either fuel. I think this works, but it adds mild complexity to the inputs/outputs. 

Are there better ways to approach this problem?

[KanORS-AssistGPT]

Fortunately, YES!
---

1. Best Option: Use `PRC_REFIT`

TIMES includes a built-in attribute called `PRC_REFIT` that does **exactly what you need** — it allows you to convert capacity from an existing process (like a coal boiler) into a new one (e.g., a biomass-capable boiler) in a given year, with an associated cost.

You just need to create a new process - for example, `Boiler_Bio-rf` - and declare `PRC_REFIT = -1` with the host process (`Boiler`) in the `other_indexes` dimension. This tells TIMES that capacity from the original boiler can be refitted into the new one.

You may also configure the new process to consume both coal and biomass, with different efficiencies, and apply upper/lower bounds on fuel shares.

---

2. Simpler Alternative: Fuel Flexibility

If you're looking for a more abstract or stylized representation:

* Let the original boiler process accept both coal and biomass as fuels.
* Use `FLO_BND` to enable biomass usage starting in a certain year.
* Reflect retrofit-related costs by applying a `FLO_DELIV` on biomass inputs.

This is useful for exploratory scenarios but doesn’t represent vintage-specific capacity shifts or explicit retrofit actions the way `PRC_REFIT` does.

---

Note: This is a great example of the advantages of advanced and mature formulation like TIMES. In MARKAL or OSeMOSYS, you’d need to manually build a workaround using auxiliary processes and flows.

[LukeSearle]

That's great! thanks very much robot friend