Reactor Materials · foundry-ml
dataset_rpv_tts
A tabular dataset of reactor pressure vessel steel measurements for predicting irradiation-induced embrittlement.
What dataset_rpv_tts is
dataset_rpv_tts is a tabular materials science dataset compiled by Ryan Jacobs, Takuya Yamamoto, G. Robert Odette, and Dane Morgan. It provides 4,535 records tracking transition-temperature shifts in reactor pressure vessel (RPV) steels.
During nuclear reactor operation, structural steel components degrade due to neutron irradiation, causing the ductile-to-brittle transition temperature to rise. This dataset supplies compositional, thermal, and irradiation parameters to support machine learning models that predict transition temperature shifts and evaluate associated uncertainties.
How it was made
The dataset consists of 4,535 entries arranged in tabular format within a single training split. Features cover steel composition by weight percent (copper, nickel, manganese, phosphorus, silicon, and carbon), exposure conditions (logarithmic irradiation fluence and flux), measurement temperature in degrees Celsius, and a data subtype descriptor. The target variable is the measured Charpy ductile-to-brittle transition temperature shift (Measured DT41J) in degrees Celsius.
Key facts
- Record count
- 4,535
- Data modality
- Tabular
- Format
- Parquet
- Target variable
- Measured DT41J [C] (ductile-to-brittle transition temperature shift)
- Input features
- Measurement temperature, wt% Cu, Ni, Mn, P, Si, C, log fluence, log flux, and datatype
- Authors
- Ryan Jacobs, Takuya Yamamoto, G. Robert Odette, Dane Morgan
- Release year
- 2023
- DOI
- 10.18126/3zkm-yd51
What you can use it for
- Transition temperature shift prediction
- Training regression models to predict ductile-to-brittle transition temperature shifts (DT41J) directly from alloy chemistry and irradiation exposure metrics.
- Embrittlement uncertainty estimation
- Benchmarking machine learning methods that provide uncertainty bounds on steel embrittlement under varying neutron flux and fluence regimes.
- Compositional sensitivity analysis
- Investigating how varying concentrations of elements such as copper, nickel, manganese, and phosphorus influence radiation-induced embrittlement.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from foundry-ml/dataset_rpv_tts at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 4,535 rows of default/train
Measured DT41J [C]
float- min
- -28.93
- median
- 39.78
- mean
- 51.84
- max
- 370.83
datatype
string_label- Plotter45%
- IVAR32%
- Plotter-15 MTR13%
log(fluence_n_cm2)
float- min
- 15.97
- median
- 19.04
- mean
- 18.91
- max
- 20.34
log(flux_n_cm2_sec)
float- min
- 8.36
- median
- 11.29
- mean
- 11.51
- max
- 14.00
temperature_C
float- min
- 250
- median
- 290
- mean
- 287.79
- max
- 315
wt_percent_C
float- min
- 0
- median
- 0.14
- mean
- 0.12
- max
- 0.34
wt_percent_Cu
float- min
- 0
- median
- 0.12
- mean
- 0.16
- max
- 0.86
wt_percent_Mn
float- min
- 0
- median
- 1.40
- mean
- 1.35
- max
- 1.98
wt_percent_Ni
float- min
- 0
- median
- 0.69
- mean
- 0.69
- max
- 1.85
wt_percent_P
float- min
- 1.5e-3
- median
- 9.0e-3
- mean
- 0.01
- max
- 0.05
wt_percent_Si
float- min
- 0
- median
- 0.22
- mean
- 0.21
- max
- 1
Strengths and limits
Good at
- Provides 4,535 experimental records linking alloy chemistry, exposure flux, and fluence to embrittlement shifts.
- Features are formatted directly for tabular regression and classification tasks in standard scientific workflows.
- Includes logarithmic scaling for neutron fluence and flux to handle wide orders of exposure magnitude.
Watch out for
- Provides only a single training split without pre-partitioned validation or test sets.
- Data coverage is restricted to the specific reactor pressure vessel steel compositions and irradiation ranges documented.
Questions
Can I look inside dataset_rpv_tts without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is dataset_rpv_tts under?
Its authors declare other. Read the terms on the original page before reusing the data.
What is the primary target variable?
The target is Measured DT41J [C], which measures the ductile-to-brittle transition temperature shift in degrees Celsius.
What input features are included?
Inputs comprise measurement temperature in degrees Celsius, weight percentages of Cu, Ni, Mn, P, Si, and C, log-scale neutron fluence, log-scale neutron flux, and a data subtype identifier.
What machine learning tasks are supported?
The dataset is annotated for both tabular regression tasks, such as predicting continuous temperature shifts, and tabular classification tasks.
How is the dataset partitioned?
All 4,535 records are supplied in a single train split.
How to cite it
Jacobs, Ryan and Yamamoto, Takuya and Odette, G. Robert and Morgan, Dane. Predictions and uncertainty estimates of reactor pressure vessel steel embrittlement using Machine learning. Materials Data Facility, 2023. https://doi.org/10.18126/3zkm-yd51
Written from dataset_rpv_tts's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.