Point of Contact: Dr. Jared Johnson
This venture under NNSA Nonproliferation Research and Development (NA-22) is focused on characterizing operations at DOE nuclear reactor and irradiated material processing testbeds through the use of data analytics on disparate data sets. To facilitate this study, the project has gathered sensor and ground truth data at the Oak Ridge National Laboratory (ORNL) Hight Flux Isotope Reactor (HFIR) and collocated Radiochemical Engineering Development Center (REDC). This section does not serve as an overview of the MINOS venture and its research efforts. Instead, this section will highlight examples of data that is collected as part of MINOS. For those interested in this data (perhaps for use in modelling and simulations), details will be presented on how to request access to this data and points of contact for collaboration with MINOS.
ETI Thrust Area 1 collaborators may be interested in MINOS data as a means for developing models and simulations that can be used in nonproliferation research. As it pertains to ETI’s mission, MINOS provides real-world data, with ground-truth for some datasets, that can confirm or inform computational models. A brief, non-inclusive list of potential datasets are listed below:
- Seismo-acoustic
- Electro-magnetic
- Biota
- Infrasound
Data is stored and managed by Lawrence Berkeley Laboratory (LBL). To begin the process, start by emailing Dr. Jared Johnson who is the MINOS venture lead. Please state that your affiliation (including your participation in ETI), your interest in collaboration with MINOS, and your intended use for the data. Upon approval, you will be asked to acknowledge and accept the MINOS Data Use Agreement. You will then be forwarded to register an account on the MINOS database and Dr. Brian Quiter of LBL will grant you access. Once this is complete you will be able to access, browse, download, and use MINOS data.
While the MINOS website has many useful tools,
this section serves to highlight a few. A great place to start for more
detailed instructions is at the
MINOS help page (found by going
to the Help
tab on the main page).
All datasets can be seen as a list by going to the Datasets
tab on the
webpage. This page allows keyword searches for specific datasets and has
various tags to sort data based on its relevance. MINOS entries can be filtered
using the organization
tag on the left-hand side of the browser. This
tag loosely groups datasets together based on the data's modality and the
institution where it was collected. Clicking on an entry will open a list of
files associated with that dataset. There should be a JSON that contains
summary information about what is contained on this page. Data is commonly
tarballed into a large database format and a more reasonably sized HDF5 format.
Clicking on any of the files brings the user to a download page for manually
downloading files. Some data (for example biota) may be in another format like
XLSX rather than those explained above.
Alternatively, data can be explored using the Data Workspace
tab on the
home page. Clicking Workspace
allows the user to select any number of
different data sets for download (remember to click Save
after choosing
datasets). Once files are in the workspace, they can be selected and then
reviewed by pressing Load Data Collections
. The bottom portion of the
page acts as a staging ground before downloading files. From here, there are
three options for the user.
Press Submit Query
and all files “checked” in the workspace will
be downloaded. The default file format downloaded will be the HDF5 datasets.
This tab allows the user to individually select HDF5 files from the selected workspace datasets for download. A valuable feature here is the ability to add a time-selection visualization. Select the time filter for the datasets and a histogram will be generated with the frequency of entries at different time bins. This may be useful in understanding when data was being collected at maximum capacity. Subsets of the data can be selected by dragging an interval over the histogram. This will highlight the selection of times to be prepared for download.
If there are any non-HDF5 files available for download in the selected workspace datasets, they can be specifically downloaded here.
If the user knows what data is needed, it can be downloaded without using the
website. A generic Rest API can be used to download
data directly to client. To start, select
Tools -> REST API Key Manager -> Generate New API Key
.
Make sure to copy this key down and take note of its expiration date
(new keys can always be generated). This section will not detail the rest of
the instructions for using the API. Instead, these instructions are provided in
the MINOS help documentation linked above.
A Python API (i.e. "MINOS API") allows the user to download and execute data
from a JupyterHub portal found on the webpage at
Tools -> JupyterHub Portal
. This is run on a cluster at LBL and allows
a user to create Jupyter notebooks for data analysis. When the Python API is
used here, data will be downloaded to the cluster, which may be more
advantageous than downloading directly to a desktop computer.