Sai dev by yvrob · Pull Request #125 · watts-dev/watts

yvrob · 2026-09-15T17:35:34Z

Description

This pull requests supports the use of Watts as part of WISE and the Systems Analysis and Integration Campaign.
This pull request adds support for Cyclus and DYMOND within WATTS and introduces a shared infrastructure for common nuclear fuel cycle analyses.

The primary additions include:

A new Cyclus plugin for executing Cyclus simulations and reading SQLite output.
A new DYMOND plugin for executing DYMOND simulations and reading HDF5 output.
Shared fuel-cycle utility modules that provide a common implementation between the two plugins for:
- reactor deployment and fleet evolution,
- generated power,
- fuel demand and discharge,
- enrichment,
- natural uranium requirements,
- tails generation,
- separative work (SWU).
Additional ACCERT result properties exposing direct, indirect, owner, and total overnight capital cost (OCC) metrics.
Registration of the new plugins within the WATTS plugin framework.
Comprehensive unit and integration tests covering the new functionality.

The shared fuel-cycle infrastructure was created to reduce duplicated logic between the Cyclus and DYMOND plugins while providing a consistent public interface for common nuclear fuel cycle analyses.

Testing

Successfully executed the Cyclus and DYMOND example workflows included in this pull request.
Added unit and integration tests covering the new plugins and shared fuel-cycle utilities.
Ran the WATTS test suite. All new tests pass. One pre-existing Windows-specific test (tests/test_fileutils.py::test_run) fails because it invokes the Unix env executable; this failure is unrelated to the changes introduced in this pull request.

Checklist

My code follows the [style guidelines](https://watts.readthedocs.io/en/latest/dev/styleguide.html)
I have performed a self-review of my own code
I have made corresponding changes to the documentation (if applicable)
I have added tests that prove my fix is effective or that my feature works (if applicable)
I have updated the CHANGELOG.md file (if applicable)
I have successfully run examples that may be affected by my changes

This reverts commit cba8004.

…sing info to output

2025 sai emanes5 See merge request system-analysis-and-integration/watts!3

…tegration/watts into cyclus_plugins

Cyclus plugins See merge request system-analysis-and-integration/watts!4

Add Cyclus plugin demo notebooks and setup script; update plugin_cyclus See merge request system-analysis-and-integration/watts!5

Dymond edits See merge request system-analysis-and-integration/watts!1

…t year

Cyclus plugin streamlining core See merge request system-analysis-and-integration/watts!22

Add documented Cyclus and DYMOND plugin tests See merge request system-analysis-and-integration/watts!23

…fuel cycle utilities as well

Co-authored-by: Ryan, Nathan <ryanns@ornl.gov>

Plugins overhaul See merge request system-analysis-and-integration/watts!24

Changed plugin See merge request system-analysis-and-integration/watts!25

yvrob · 2026-09-15T17:36:52Z

@nstauff, here are the additions to Watts we developed for the SA&I campaign, to be used with WISE. I would appreciate if we could work together to make this mergeable if any issue comes up. Thank you!

nstauff · 2026-09-15T18:10:59Z

@abachma2 and @JiaZhou-PU - Could you add yourself as reviewer and confirm you are OK with this PR?

abachma2 · 2026-09-15T18:16:46Z

I don't see a way to add myself as a reviewer -- I don't see that option. I will review this later this week, likely on Friday. My review will focus on the Cyclus and DYMOND additions/changes, Jia can speak more to the ACCERT changes. Feel free to ping me if you don't see anything by next week.

JiaZhou-PU · 2026-09-15T18:20:33Z

@nstauff I am good with those changes in ACCERT, although I will submit another PR based on the recent ACCERT version.

JiaZhou-PU · 2026-09-15T18:29:02Z

+        """Total OCC = Total cost without owner + Owner cost"""
+        return (self.account_table['total_cost'].values[0] * 1.609 * 1.2 /
+                0.834)
+


This is the default PWR algorithm for the OCC in ACCERT. I will expand on those later

nstauff

Thanks - see my submitted comments. Flagging for @paulromano to complete a review and discuss how best organize the tests and utilities within WATTS.

nstauff · 2026-09-16T16:06:21Z

+# SPDX-License-Identifier: MIT
+
+"""
+This example demonstrates how to use WATTS to perform a criticality search on an


please update the description here - this appears to be a bad copy/paste

nstauff · 2026-09-16T16:06:50Z

+params = watts.Parameters()
+
+
+# Create MCNP plugin


please update comment. - this is cyclus plugin...

nstauff · 2026-09-16T16:11:12Z

    # Check for environment variable
    base_dir = os.environ.get(environment_variable)
-    return base_dir / exe if base_dir is not None else exe
+    if base_dir is None:


I am not sure to understand what those changes to plugin do, and what motivates those. Flagging for @paulromano review.

nstauff · 2026-09-16T16:14:07Z

@@ -0,0 +1,427 @@
+import numpy as np


@paulromano - please let me know how you want this watts source code to include additional utility files.

nstauff · 2026-09-16T16:21:18Z

@@ -0,0 +1,211 @@
+import pytest
+import numpy as np


I recommend renaming the test to make it more explicit - this is a test on utility "enrichment" features. Some description/comments would be useful as well.

I recommend changing the file name to test_enrichment_utils.py to make it more clear that you are testing the functions in the enrichment_utils.py file.

abachma2

I'm still working through the files, it's taking me longer than expected, but I wanted to get you some feedback in the mean time. A few major items:

Two tests fail for me locally: tests/test_fileutils.py, test_run (only on my Windows machine) and tests/test_parameters.py, test_parameters_show_summary (on my Windows and Linux machines).
I recommend going through the comments you have and seeing what is necessary. The code I have seen is quite cleanly written, so comments should be more about why you are doing something and less on what the code is actually doing. I flagged a few cases of this.
It seems like the cyclus and dymond examples are quite different -- they even have a different number of variable parameters in the template. These codes (for the purposes of a versatile tool like WATTS) do the same things and it might make sense for them to have similar examples.

I'll keep working on the rest of my review next week.

abachma2 · 2026-09-18T18:21:24Z

+        <fuel_outrecipes> <val>lwr_used_fuel_recipe</val> </fuel_outrecipes>
+        <fuel_incommods>  <val>enriched_u</val>           </fuel_incommods>
+        <fuel_outcommods> <val>waste</val>                </fuel_outcommods>
+        <fuel_prefs>      <val>0.5</val>                  </fuel_prefs>


What is the purpose of having this fuel pref?

abachma2 · 2026-09-18T18:23:08Z

+    <nuclide> <id>942410000</id> <comp>36.477</comp> </nuclide>
+    <nuclide> <id>942420000</id> <comp>15.387</comp> </nuclide>
+    <nuclide> <id>952410000</id> <comp>1.234</comp> </nuclide>
+    <!-- <nuclide> --> <!--   <id>95242m</id> --> <!--   <comp>0.03</comp> --> <!-- </nuclide> -->


Any reason this nuclide is commented out?

abachma2 · 2026-09-18T18:23:57Z

@@ -0,0 +1,18 @@
+# 1App_MCNP_Jezebel


Suggested change

# 1App_MCNP_Jezebel

# 1App_Cyclus

abachma2 · 2026-09-18T18:24:19Z

+
+## Purpose
+
+This example provides a demonstration on how to use WATTS to run Cyclus.


Should you add a description of what the example problem is?

abachma2 · 2026-09-18T18:31:34Z

+## File descriptions
+
+- [__watts_exec.py__](watts_exec.py): WATTS workflow for this example. This fills the duration of the Cyclus run.
+- [__cyclus_template.xml__](cyclus_template.xml): Cyclus input file with duration template.


Suggested change

- [__cyclus_template.xml__](cyclus_template.xml): Cyclus input file with duration template.

- [__cyclus_template.xml__](cyclus_template.xml): Cyclus input file template with simulation duration as a variable.

This wording feels more clear to me.

abachma2 · 2026-09-18T20:16:21Z

+    for n in range(n_stages - 1, -1, -1):
+        swu = qty_prod[n] * swu_to_product_ratio(
+            feed_enrs[n], prod_enrs[n], tails_enrs[n]
+        )


In this loop it feels like you have some variable assignments that maybe aren't necessary. LIke maybe move the swu calculation to the end and assign it directly into the swu_list variable, and then have the printing at the very end.

abachma2 · 2026-09-18T20:17:21Z

+    )
+    cascade_data = {}
+    for key in multistage_dict:
+        cascade_data[key] = list(multistage_dict[key])


Isn't this already a list?

abachma2 · 2026-09-18T20:23:46Z

+
+# DYMOND process times use a Julian year.  Defining it here keeps both plugins
+# on the same time conversion when they report common WATTS results.
+SECONDS_PER_YEAR = 365.25 * 24 * 60 * 60


Maybe something later on will help elucidate how time is handled in all of this, but:

This value doesn't match the number of seconds per year in Cyclus when using the default time step. The default time step in Cyclus is 2,629,46 seconds, or 1/12th of a year and that times 12 does not equal this variable.

Cyclus allows for a variable time step size -- do you have a way to account for that in the processing?

abachma2 · 2026-09-18T20:30:52Z

+    Attributes
+    ----------
+    obj : CyclusOutput
+        Reader for the Cyclus SQLite file.  The file is opened when this


Cyclus can (in theory) yield results databases in other formats, including HDF5. Does this class only work with SQLite databases? If so, that should be documented.

abachma2 · 2026-09-18T20:31:45Z

+        # WATTS moves the SQLite file after this object is created.  Open the
+        # file only when a caller first requests ``obj`` so the reader uses the
+        # final path and WATTS never tries to save an open SQLite connection.


Does this need to be here, or should it be in the docstrings?

abachma2

I finally got a chance to look through everything. Thanks for your patience. A few big items/themes:

I would check on the vocabulary used in each plugin and make sure it matches the correct vocabulary for each code. It might confuse a user if vocabulary isn't consistent with what each code uses.
There are multiple places where the first part of a function looks to be changing the datatype of one of the inputs, and it isn't clear to me why you would do that instead of just requiring the parameter to be of that datatype.
It looks like there are places in which you want the Cyclus and DYMOND plugin to match perfectly, which sometimes leads to mixing of vocabulary. Do the plugins need to match perfectly? Can they have functions that do the same thing, but are called different things because of different vocabularies?
I can run the Cyclus example locally, but I can't get the DYMOND example to run. It looks like it creates the input file correctly, but them dymond can't seem to run the file and I can't know why. I'll keep looking into this one to see if I can figure anything else out.
Is there any reason you didn't update the CHANGELOG file for this PR?

abachma2 · 2026-09-21T13:39:54Z

+def find_single_output(
+    outputs: Iterable[Path],
+    suffixes: Sequence[str],
+    description: str,
+) -> Path:


This function is meant to find if an output file exists, and get that file path, correct? As written, it feels too complicated for this task. Can you provide some context/information on why it's written this way?

abachma2 · 2026-09-21T13:44:19Z

+    return str(unit).strip().upper().replace(" / ", "/")
+
+
+def power_to_gw(values, unit: str) -> np.ndarray:


Any reason you selected GW as the unit to convert things to? As opposed to MW?

abachma2 · 2026-09-21T13:45:44Z

+        If a requested reactor is not present in ``available``.
+    """
+
+    available_names = tuple(str(name) for name in available)


Why not just require that available be defined as a tuple? That would remove the need to define another variable and to assign a new datatype to it.

abachma2 · 2026-09-21T13:51:44Z

+    # A dictionary preserves insertion order, so it removes repeated names
+    # without changing the order supplied by the caller.
+    requested = tuple(dict.fromkeys(requested))


I thought dictionaries were unordered, which is why you can't get a value based on index. So I'm a little confused on the purpose of this part.

abachma2 · 2026-09-21T14:20:44Z

+def aggregate_calendar_years(
+    values,
+    init_year: int,
+    init_month: int,
+    how: str,
+) -> tuple[np.ndarray, np.ndarray]:
+    """Combine monthly values by calendar year.


Does this function have baked into it the idea that each timestep is 1 month? Does it allow any flexibility for if a timestep isn't 1 month?

abachma2 · 2026-09-23T19:20:45Z

+        Parameters
+        ----------
+        reactors : str or list[str]
+            Prototype name(s).


I don't think dymond uses "prototype" in it's vocabulary. Is there another term that can be used here?

abachma2 · 2026-09-23T19:23:46Z

+
+        return collected
+
+# =================================DEPRECATED==================================


Do we need to keep the deprecated functions?

abachma2 · 2026-09-23T19:24:32Z

+    def _normalize_facilities(
+        self,
+        facilities,
+        facility_class="reactor",


Does it make sense to have a default value for this parameter?

abachma2 · 2026-09-23T19:25:35Z

+@pytest.fixture
+def dymond_output(tmp_path):
+    """Create a compact but representative DYMOND HDF5 output.


Would it be easier to just store an output database for testing?

abachma2 · 2026-09-23T19:30:35Z

+        cascade_splits=[],
+    )
+
+    cascade_5 = enrichment_cascade(


Are you using another function to test the output of this one? Does the one function use the enrichment_cascade function?

abachma2

I finally got a chance to look through everything. Thanks for your patience. A few big items/themes:

I would check on the vocabulary used in each plugin and make sure it matches the correct vocabulary for each code. It might confuse a user if vocabulary isn't consistent with what each code uses.
There are multiple places where the first part of a function looks to be changing the datatype of one of the inputs, and it isn't clear to me why you would do that instead of just requiring the parameter to be of that datatype.
It looks like there are places in which you want the Cyclus and DYMOND plugin to match perfectly, which sometimes leads to mixing of vocabulary. Do the plugins need to match perfectly? Can they have functions that do the same thing, but are called different things because of different vocabularies?
I can run the Cyclus example locally, but I can't get the DYMOND example to run. It looks like it creates the input file correctly, but them dymond can't seem to run the file and I can't know why. I'll keep looking into this one to see if I can figure anything else out.
Is there any reason you didn't update the CHANGELOG file for this PR?
It looks like the license text on some of the files says 2024 and others say 2026. Any reason for the difference?

jbae11 and others added 30 commits February 24, 2025 14:39

flows, stocks, and products

cba8004

Revert "flows, stocks, and products"

eaf1949

This reverts commit cba8004.

Adding DYMOND plugin and files for FY24 scenario

c80b659

update to discretize by prototype, not agentid

697fa94

Added placeholder helper functions to the results and needed reproces…

18380ae

…sing info to output

Switched stocks file to one with fabrication losses and correct waste

9f3e6d2

errors when there's no fuel for reprocessed reactors

dcbc805

example of using dymond plugin api to get mass flows

fff9b5d

input edit to be dict and not list of dict

4f49604

2025 sai emanes5

b8d0803

Merge branch '2025_sai_emanes5' into 'cyclus_plugins'

cabe10f

2025 sai emanes5 See merge request system-analysis-and-integration/watts!3

merge conflict

36564db

2025 sai emanes5

e32fbba

Merge branch 'cyclus_plugins' of code.ornl.gov:system-analysis-and-in…

8ac19e4

…tegration/watts into cyclus_plugins

Merge branch 'cyclus_plugins' into 'development'

1c51c40

Cyclus plugins See merge request system-analysis-and-integration/watts!4

add natu and swu calculations

aacc1b2

add fuel example

d35a5ff

better examples

15dd6e4

cleanup

24cf8e1

test for adding vogtle as not legacy

089062b

Add Cyclus plugin demo notebooks and setup script; update plugin_cyclus

9111868

Merge branch 'cyclus_update' into 'cyclus_plugin'

8545bcd

Add Cyclus plugin demo notebooks and setup script; update plugin_cyclus See merge request system-analysis-and-integration/watts!5

multistage enr

dbc3890

accept incoming

169da26

Add back split_cat enrichment

bc77e55

Fixed enrichment to U denominator instead of total mass

0ce802d

Merge branch 'dymond_edits' into 'dymond_plugin'

f8bbdf5

Dymond edits See merge request system-analysis-and-integration/watts!1

Replaced example files with what should be the final version used las…

379fb7e

…t year

Better logging for enrichment

77d46d3

Better enrichment function, faster proto ID scraping

c8817a8

Robert and others added 16 commits June 9, 2026 20:53

Add Cyclus capacity factor time series

7fc5dda

Merge branch 'cyclus_plugin_streamlining_core' into 'development'

c208a6b

Cyclus plugin streamlining core See merge request system-analysis-and-integration/watts!22

Add documented Cyclus and DYMOND plugin tests

e903bfc

Merge branch 'cyclus-dymond-documented-tests' into 'development'

845aa2b

Add documented Cyclus and DYMOND plugin tests See merge request system-analysis-and-integration/watts!23

Clarified multistage enrichment

c7e9d29

Added common utility function for fuel cycle simulators

52e54fa

Integrated fuel cycle utility function and fixed caching

e0a1ea4

Integrated fuel cycle utility function in DYMOND plugin

35a328a

Added tests for cyclus and dymond plugins, common tests for both and …

b0b9dfa

…fuel cycle utilities as well

Included exception for product assay < nat assay (DU/recycling)

4b10edb

Fixed path issue with Windows and unix for plugin

8e4c32b

Apply 1 suggestion(s) to 1 file(s)

f76f0db

Co-authored-by: Ryan, Nathan <ryanns@ornl.gov>

Added default assays and cascade split

4221de1

Merge branch 'plugins_overhaul' into 'development'

bba64b3

Plugins overhaul See merge request system-analysis-and-integration/watts!24

Merge branch 'changed_plugin' into 'development'

52be402

Changed plugin See merge request system-analysis-and-integration/watts!25

Merge upstream development into SAI development

171dc21

nstauff assigned nstauff and paulromano Sep 15, 2026

abachma2 self-requested a review September 15, 2026 18:26

JiaZhou-PU self-requested a review September 15, 2026 18:27

abachma2 added the enhancement New feature or request label Sep 15, 2026

JiaZhou-PU approved these changes Sep 15, 2026

View reviewed changes

nstauff requested changes Sep 16, 2026

View reviewed changes

abachma2 requested changes Sep 18, 2026

View reviewed changes

abachma2 requested changes Sep 23, 2026

View reviewed changes


		## Purpose

		This example provides a demonstration on how to use WATTS to run Cyclus.

	- [__cyclus_template.xml__](cyclus_template.xml): Cyclus input file with duration template.
	- [__cyclus_template.xml__](cyclus_template.xml): Cyclus input file template with simulation duration as a variable.

		return str(unit).strip().upper().replace(" / ", "/")


		def power_to_gw(values, unit: str) -> np.ndarray:


		return collected

		# =================================DEPRECATED==================================

Conversation

yvrob commented Sep 15, 2026

Description

Testing

Checklist

Uh oh!

yvrob commented Sep 15, 2026

Uh oh!

nstauff commented Sep 15, 2026

Uh oh!

abachma2 commented Sep 15, 2026

Uh oh!

JiaZhou-PU commented Sep 15, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nstauff left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

abachma2 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

abachma2 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

abachma2 left a comment

Choose a reason for hiding this comment

Uh oh!