Feature Request: Multiple Construction Years and/or non-Year 0 Capital Expenditures

[softwareengineerprogrammer]

Is your feature request related to a problem? Please describe.

Yes - we have a problem with being able to accurately calculate NPVs/IRR in GEOPHIRES for geothermal projects.

Geothermal projects are typically expected to have multi-year permitting, exploration, and construction timelines. The GEOPHIRES SAM Economic Model is currently only able to support 1 construction year due to SAM's financial model allowing multiple construction/pre-revenue years and/or non-Year 0 capital expenditures.

Describe the solution you'd like

Support either or both of:

1. Multiple pre-revenue years (before Year 0) that allow per-year allocations of capital expenditure
2. Support capital expenditures after Year 0

Describe alternatives you've considered

1. In some cases it is acceptable to calculate time-adjusted NPVs from SAM cash flow revenue/returns line items, such as royalty streams, but this is not adequate for the tax investor's NPV.
2. Additional capex can be added to the installation cost passed to SAM (similar to the SAM UI's Construction Financing control), but this is also only a very rough approximation that is not appropriate for a proper cash flow analysis.

Additional context

Happy to discuss our use case in detail - please follow the links in my profile (@softwarengineerprogrammer) for contact info (or we can discuss on this issue, whatever works best for you).

Thanks for your help and consideration!

[cpaulgilman]

Thank you for this feature request. SAM's financial models currently assign all development and construction-related costs to Year 0 of the project cash flow.

We do not have immediate plans or funding to add costs for a multi-year development/construction period, but this is a limitation we are aware of and would like to address.

[softwareengineerprogrammer]

Thank you for this feature request. SAM's financial models currently assign all development and construction-related costs to Year 0 of the project cash flow.

We do not have immediate plans or funding to add costs for a multi-year development/construction period, but this is a limitation we are aware of and would like to address.

Thanks @cpaulgilman. It's possible I may wind up implementing this myself in a fork/extension. Do you have any design documents or thoughts on architecture/approach that might help me steer in a direction you/NREL are aligned with?

Feel free to contact me via contact methods linked in my profile to take the discussion offline, if preferred.

[brtietz]

@softwareengineerprogrammer Thrilled to hear you're interested in this. To clarify, we do have some budget to bring in open-source contributions like this one, just not to implement it ourselves. If you make a fork, please make a PR and we'll work with you to get it in the upstream version, as well as answering questions like you're asking now.

As far as NREL aligned implementations, are you familiar with the approach taken in the Electricity ATB? See the Geothermal and Nuclear tabs of https://data.openei.org/files/6006/2024_v3_Workbook.xlsx for details.

[softwareengineerprogrammer]

Thanks for the guidance @brtietz. I now have a reasonably well-featured prototype implementation cooking at softwareengineerprogrammer/GEOPHIRES#112. (I can give you a nudge once it's fully fleshed out as a release-worthy PR into the main GEOPHIRES repo, if you're interested.)

The cruxes of the implementation are:

1. A pre-processor to compute pre-revenue cash flows and then pass time-, interest-, and inflation-adjusted installation cost and debt balance to the SAM financial engine
2. A post-processor that minimally alters the SAM financial engine cash flow result by:
   1. Inserting CONSTRUCTION category line items for pre-revenue (construction) years
   2. Backfilling IRR & NPV with combined pre-revenue and operational phase cash flows

I'd be more than happy to discuss further if you feel there might be value in using my approach as a reference. I would also very much welcome design and/or code review/feedback (formal or informal).

[brtietz]

@softwareengineerprogrammer Thanks for sharing! This looks like a good start. A couple of comments/questions:

1. Can you explain more about what's in the inflation cost line? I'm having a bit of trouble telling what's cumulative and what's annual. The fact that the year 0 inflation as a percentage of overnight costs is larger than year -6 feels off to me, but I might not understand what you're trying to do there (is this present value or nominal dollars?)
2. ATB represents different costs of capital (discount rates) pre-drilling and post-drilling. Is that something you're planning on adding?

[softwareengineerprogrammer]

Thanks for the feedback @brtietz!

1. Can you explain more about what's in the inflation cost line? I'm having a bit of trouble telling what's cumulative and what's annual. The fact that the year 0 inflation as a percentage of overnight costs is larger than year -6 feels off to me, but I might not understand what you're trying to do there (is this present value or nominal dollars?)>

Everything in that screenshot is annual - I can see how it might look cumulative but that's just an artifact of the phased CAPEX schedule ramping up year over year until operations start. I will take that as a point to clarify in output and/or documentation.

2. ATB represents different costs of capital (discount rates) pre-drilling and post-drilling. Is that something you're planning on adding?

That is a consideration that has recently popped up on my radar - I wasn't planning on it for the initial implementation, but your question lends more weight to its prioritization.

Truly appreciate the feedback - as I mentioned previously, I'll nudge you when I have a more reviewable version ready (though of course, any additional feedback in the interim is more than welcome as well).

[softwareengineerprogrammer]

@brtietz Here is the review-ready version of the PR, as promised: NREL/GEOPHIRES-X#440. (I've marked it as Draft for now until I can get some nominal SAM SME initial review, such as from yourself.)

You'll probably be most interested in the documentation writeup (also linked from the PR description): https://softwareengineerprogrammer.github.io/GEOPHIRES/SAM-EM_Multiple-Construction-Years.html

It contains the following design diagram and screenshots, included below for quick reference:

[brtietz]

Thanks for sharing. What is the discount rate in the above example?

For a real discount rate higher than 0, I would expect the after tax cumulative NPV line in year -6 to be larger than the nominal cash flow to account for the time-value of those 6 years.

Unfortunately the equations at https://atb.nrel.gov/electricity/2024b/equations_&_variables don't explain this very well, but you can see the effect in the ATB spreadsheet (https://atb.nrel.gov/electricity/2024b/data) by shifting the construction costs from debt to equity. The construction finance factor increases with more equity since equity is typically more expensive than debt. I'd expect to see the same effect here on NPV. (I'll try to get these equations updated for the next ATB release) Appendix F of https://docs.nrel.gov/docs/legosti/old/5173.pdf has equations for these effects.

Once we get aligned on discount rates, I'll do a deeper dive on the unit tests in https://github.com/softwareengineerprogrammer/GEOPHIRES/tree/lcoe-nominal-derived. Happy to serve as that SME, feel free to add me as an assignee on that issue.