DYMOND is an agent-based, event-driven, fuel cycle simulator. It can model time-dependent behavior of evolving fuel cycles, where new facilities and technologies are added to the supply chain and outdated technologies are retired. Mass flows and isotopics are udpated on-the-fly during mining, enrichment, fabrication, reactor burnup, cooling, reprocessing, and disposal. Flexible model construction allows for unlimited facility or reactor types, multiple levels of fidelity, and any number of nuclides to be tracked.
DYMOND is awaiting authorization for public release.
Once authorized, links to download the code and user manual will be added here.
About Version 7
The Dynamic Model of Nuclear Deployment (DYMOND) code is a nuclear fuel cycle simulator (NFCS) first developed as a system dynamics model at Argonne National Laboratory in 2001. Since then, it has undergone significant development which has added capabilities and altered the underlying modeling paradigm to improve the flexibility and accuracy of the code. The latest release, Version 7, uses a fully agent-based and event-driven modeling paradigm.
Benefits of being agent based – Modular architecture simulates the nuclear fuel cycle as a network-graph-like system where materials and messages are passed through direct agent-to-agent connections. Global or macro-level properties are from emergent behavior of agent interactions rather than dictated characteristics. Flexible model construction allows for unlimited facility or reactor types, multiple levels of fidelity, and any number of nuclides to be tracked.
Benefits of being event driven – Being event–driven, the model is in continuous time (no time discretization or synchronization errors), and all state updates are on-the-fly, avoiding computationally expensive extraneous calculations.
How to Use DYMOND
Step 1: Build the Fuel Cycle
In Version 7, users build fuel cycles by (1) adding facilities, (2) specifying the process physics occurring within each facility, and (3) connecting facilities with supply-chain relationships. DYMOND allows for complex user inputs, including initial fleet capacity, facility deployment and retirement, evolving reactor and fuel technologies, fluctuating energy demand, recycling strategies, fuel use priorities, and simulation duration. DYMOND models the nuclear fuel cycle as a system of facilities and storages. It supports 6 explicit classes of facilities and storages: enrichment facilities, fuel fabrication facilities, reactors, reprocessing facilities, generic facilities, and storages. Several of these classes also support multiple process models.
Step 2: Run the Fuel Cycle Simulator
Fuel batch movement between agents (e.g., power plant refueling) are handled by an agent-to-agent message-handling system. This relational system is defined by the user in the form input and output connections between fleets of facilities. Simulations are driven by events scheduled either by the user or by agents. Agent life-cycle and process models automatically schedule the majority of events, but user-defined events allow for changes in agent behaviors and characteristics, model states, and materials inventories and movements external to the model. Material separations and transmutations can be handled via recipes, relational factors, or depletion and decay calculations. Interfaces for external tools depletion tools, such as ORIGEN-S and OpenMC.deplete, can also be used for criticality and burnup calculations for advanced fuel fabrication searches.
Step 3: Output Results
DYMOND’s HDF5 output can be integrated with other modeling codes including: DAKOTA, A-LEAF, GCMat, ACCERT, and WATTS. (More details on these codes can be found below.)
Standard Code Details
This development of this code has been sponsored by the U.S. Department of Energy, Office of Nuclear Energy’s Systems Analysis and Integration Campaign.
System Requirements: Any OS with Java JDK 22+ installed.
Machine Requirements: It is recommended to have 4 GB or more memory available (designed for 8GB) and more than 2 GB free disc space for output files.
File Types: JSON input and output summary files, HDF5 standard output files.
Typical Runtime: Run times are highly variable. Simple problems using recipes will run in a few seconds. Whereas a simulation of hundreds of years with thousands of agents and physics calculations could take several hours.
Auxiliary Codes: ORIGEN-S, OpenMC
DYMOND v7 is implemented in the java programming language and now supports JavaScript Object Notation (JSON)–formatted input files and HDF5 database outputs. Optional JSON output summary files can also be created via DYMOND’s built-in postprocessing capabilities. Compared with the previous Excel-based outputs, the HDF5 format is more compact, has faster read and write times, and is more easily parsed. These developments have allowed an expansion of the DYMOND modeling ecosystem, which includes integration with:
- DAKOTA (Design Analysis Kit for Optimization and Terascale Application)
- A-LEAF (Argonne Low-carbon Electricity Analysis Framework)
- GCMat (Global Critical Materials agent-based model)
- ACCERT (Algorithm for the Capital Cost Estimation of Reactor Technologies)
- WATTS (Workflow and Template Toolkit for Simulation)