PRISM: Probabilistic Remaining-life Inference for Structural Materials is an open-source Python framework that implements the five-step uncertainty quantification (UQ) workflow for probabilistic prediction of remaining useful life (RUL) in structural materials. It delivers robust, statistically-defensible predictions of RUL for components governed by creep, fatigue, or other time-dependent degradation mechanisms.
- Overview
- Features
- Usage
- Documentation
- Reproducibility
- Authors
- Citing PRISM
- License
- Institutional support
- Funding
PRISM targets engineers and scientists who must characterise lifetime variability instead of quoting a single deterministic “safety factor”.
The default implementation ships with three classical time–temperature–parameter (TTP) creep models – Larson-Miller, Orr-Sherby-Dorn and Manson-Succop – wrapped in a uniform probabilistic interface. Because the code is modular, any analytical or data-driven lifetime model can be plugged into the pipeline. The underlying results are reported in the following publication:
- V.V. Maudonet, C.F.T. Matt and A. Cunha Jr, A framework for probabilistic prediction of remaining useful life in structural materials, 2025 (under review)
Preprint available here.
- End-to-end UQ pipeline for lifetime prediction (creep & fatigue ready)
To get started with PRISM, follow these steps:
- Clone the repository:
bash git clone https://github.com/americocunhajr/PRISM.git
- Navigate to the code directory:
bash cd PRISM/PRISM-1.0
- Execute the main file:
bash XXX
The routines in PRISM are well-commented to explain their functionality. Each routine includes a description of its purpose, inputs, and outputs.
Simulations done with PRISM are fully reproducible, as can be seen on this CodeOcean capsule.
- Victor Vieira Maudonet (UERJ)
- Carlos Frederico Trota Matt (CEPEL)
- Americo Cunha Jr (LNCC / UERJ)
We ask the package users to cite the following manuscript in any publications reporting work done with our code or data:
- V.V. Maudonet, C.F.T. Matt and A. Cunha Jr, A framework for probabilistic prediction of remaining useful life in structural materials, 2025 (under review)
@article{Maudonet2025PRISM,
author = {V.V. Maudonet and C.F.T. Matt and A. {Cunha~Jr}},
title = {A framework for probabilistic prediction of remaining useful life in structural materials},
journal = {Under Review},
year = {2025},
volume = {~},
pages = {~},
doi = {~},
}
PRISM is released under the MIT license. See the LICENSE file for details. All new contributions must be made under the MIT license.
