@ahawkes @alexdewar I wonder if we could have the following parameters for each process:
capacity_granularity: amount of capacity to appraise/add at a timedivisible (or is_divisible): a boolean flag for whether or not the asset is divisible. If divisible, the capacity_granularity chunks of capacity effectively represent distinct real-world objects (or groups of objects) which can be decommissioned independently, whereas for non-divisible assets, capacity_granularity chunks are combined into a single object with an overall capacity, which must be decommissioned all in one go
Could take many appraisal rounds if users set capacity_granularity very small compared to overall capacities, but I think that can be on the user. At least the name kind of implies a link to performance/precision. Could maybe have a "sensible" default of 1/capacity_to_activity, or something.
I think we also need to think about the issue raised in #1319 (comment). For divisible assets we may be fine as it is (even then I'm not sure - I guess it depends whether the individual objects are coordinated (e.g. wind turbines in a wind farm) or not (e.g. gas boilers in peoples homes)), but for non-divisible assets maybe we need to add capacity chunks to the growing asset at the appraisal stage, appraise the full capacity, and find a way to calculate the marginal LCOX/NPV of that extra capacity (if that even makes sense?).
Originally posted by @tsmbland in #1199
Originally posted by @tsmbland in #1199