MUSE XIII engagement

[alexdewar]

MUSE XIII engagement

This discussion will be used as an agenda for the meeting so we can determine which issues should be tackled for this upcoming engagement.

Summary of previous engagement

• Discussion page: MUSE XII engagement #918
• Previous proposal (private link)
• See release notes
• List of closed issues (partial)

This has been a highly productive engagement. After most of the basic functionality was put in place in v2.0.0, we have been able to steadily add new features as well as fixing bugs and peculiar edge cases (in particular, those not previously covered by example models).

From the proposal, there is only one issue marked as "must have" which was not completed (#523), as it was ultimately deemed less important to the model. Besides that, the circularities feature -- also a "must have" -- largely works, though not currently for every model (see discussion: #1089).

There were a few "could haves" which were not attempted:

• Add support for flexible assets #360
• Add commodity trade parameters #509
• Allow for adding commodity constraints #143

There were many updates to the model (and the program, infrastructure etc.) made which were not included in the proposal.

Aside from the many model-related improvements, there are a couple of infrastructure improvements I wanted to highlight:

• We added a new repo for MUSE2-related data analysis. It currently only has a couple of example analyses, but we are planning on expanding it as we go along (see discussion).
• Versions of documentation for old releases are now available on the documentation page. This will be useful for when we make more releases.

Plans for current engagement

Improvements to model suggested by @ahawkes:

• Change the way LCOX appraisal tool works #1199
• Change the way the NPV appraisal tool works #1201

And possible bugs:

• Investment appears to decommission too much in this case #1131
• Model is infeasible, yet there is no unmet demand (strict availability constraints) #1132
• LCOX and NPV tools do not recalculate candidate asset capacities between tranche assessments #1202
• Capacity requirement calculation might be incorrect for seasonally and annually-balanced commodities #1203

From previous engagement:

• Add support for flexible assets #360 (@tsmbland says needed for MUSE UK)
• Add commodity trade parameters #509
• Allow for adding commodity constraints #143

Our suggestions:

• Parallelise investment step #1101
• Improve treatment of divisible assets #1061
• More unit tests would be good (e.g. Add unit tests for dispatch optimisation #448)
• IMPORTANT (possibly more of a research question?): Cannot use "marginal" or "full" pricing strategies for markets with circularities #1168
• IMPORTANT: Potential for different dispatch results on different platforms #1174 and Regression test disabled on macos #1197

Infrastructure (for example):

• Tidy-ups for "old docs" scripts #1187
• Use proper diff algorithm for output of failing regression tests #1186
• Generate debug output files for all regression tests, even if they're not checked #1198

Other questions

...much the same as last time!

• How much time do we have? Note the RSE team's charge-out rate has increased.
• How much availability will @tsmbland have? (We said ~50% for previous engagement -- not sure how that worked out in practice?)
• Do we want to schedule regular catch-ups or are we happy communicating asynchronously?