STMutants: 110 Mutants of IEC 61131-3 Structured Text Programs

110 generated first-order mutants derived from 11 Structured Text programs used in industrial automation, created by Md Humaun Kabir and released in 2026. The dataset was constructed through a systematic four-phase methodology and includes mutants covering seven operator categories adapted for the PLC domain. It serves as a reproducible benchmark for research on automated test generation and mutation analysis for safety-critical PLC software.

Use Cases

• Benchmarking automated test generation tools based on the provided set of 108 observable mutants.
• Evaluating mutation analysis techniques for PLC software using the defined seven mutation operator categories.
• Training or assessing large language models for fault detection based on the baseline evaluation of GPT-4.2, Gemini 2.5 Pro, and Claude Sonnet 4.5.
• Researching fault localization in safety-critical industrial software based on the covered fault classes like value, condition, computation, and state initialization.

Strengths

• 108 mutants retained after rigorous observability and equivalence screening, with strong inter-rater agreement (Cohen's κ = 0.87).
• Covers seven defined mutation operator categories adapted for the PLC domain.
• Includes programs from both an open-source library (OSCAT) and industrial sources, spanning 38 to 211 lines of code.
• Provides a baseline evaluation of three large language models, with reported mutation detection accuracies.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is small in scale (208.5 KB), which may limit suitability for large-scale machine learning training.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Mutants were generated through a systematic four-phase methodology encompassing fault-type profiling, syntactic transformation, compilability verification, and manual equivalence screening.

Freshness

Last updated 2026-04-26 16:18:05; freshness should be verified.