FABEST LCF Benchmark Dataset: Cyclic Softening, Ratcheting and Mean Stress Relaxation of 42CrMo4+QT Steel

Overview

This dataset provides the high-quality experimental basis for the FABEST Low-Cycle Fatigue (LCF) Competition, an international benchmark study organized under the umbrella of the COST Action CA23109 FABER (Fatigue Benchmark Repository), specifically within Working Group 4.7 (Low-Cycle Fatigue).

The benchmark focuses on the complex cyclic behavior of 42CrMo4+QT steel (high-strength chrome-molybdenum alloy). It is designed to challenge the current state-of-the-art in computational mechanics by providing a rigorous "blind prediction" framework.

Scientific Objectives

The primary goal of this benchmark is to assess the capability of modern constitutive models and fatigue criteria to simultaneously capture several complex material phenomena using a single, unified set of parameters.

The competition is designed to address two key research questions:

·     Is the current generation of advanced cyclic plasticity models capable of describing both ratcheting and mean stress relaxation with a single, unified set of material parameters?

·     What is the most suitable fatigue life prediction method for uniaxial loading that can accurately account for the combined effects of both these phenomena?

Dataset Description

The provided data was obtained from a comprehensive experimental campaign conducted at VŠB – Technical University of Ostrava using Biaxial hydraulic testing machine LabControl 100kN/1000Nm. The dataset is divided into two main parts:

Calibration Set (Cases 1, 2, and 3): Full loading histories and cycle-by-cycle summaries for 24 uniaxial LCF tests (strain-controlled Rϵ=−1, strain-controlled with mean strain of 0.5 %, and stress-controlled ratcheting tests with R=−0.5). These are intended for model parameter identification and tuning.

Validation Set (Cases 4 and 5 - Blind Prediction): Loading definitions for two new scenarios (increased mean strain and different stress ratio). The experimental response for these cases is withheld to ensure a true blind prediction challenge.

Data processing has been standardized using open-source Python toolsets developed within the FABER network to ensure maximum consistency and interoperability.

Structure of the Repository

The downloadable ZIP archive includes:

Experimental Protocols: CSV files with high-resolution time-history data and stabilized cycle summaries.

Official Task Description: A PDF document outlining the competition rules, mandatory submission subsets (17 experiments), and evaluation metrics.

Submission Templates: Standardized CSV formats for Category A (Cyclic Plasticity) and Category B (Life Prediction).

Reference Paper: A pre-print of the EAN2026 conference paper detailing the experimental procedures and material characterization.

How to Participate

Participation is open to researchers from both academia and industry.

Download the dataset and the Official Rules and Task Description PDF.

Submit your predictions via the official submission portal (link available on the FABER website after the Training School in July 2026).

Winners will be announced and awarded during the Fatigue 2027 conference in Cambridge, UK.

Acknowledgement

This work was supported by the COST Action CA23109 FABER, supported by COST (European Cooperation in Science and Technology) and by the Specific Research Grants under the registration no. SP2025/048 and SP2026/027 (Department of Applied Mechanics, Faculty of Mechanical Engineering, VŠB – Technical University of Ostrava).

Contact: Prof. Radim Halama ([e-mail]) – Lead of WG 4.7 LCF.Official Website: https://www.faber-cost.eu/