Large Models Empowered Fault Detection and Diagnosis Method for Advanced Equipment

Traditional fault detection and diagnosis (FDD) pipelines often depend on handcrafted features, narrow domain models, and fixed assumptions about operating conditions. In contrast, modern engineered equipment couples sensing, control, actuation, computation, communications, and decision-making, where faults may appear as subtle cross-subsystem interactions rather than isolated component failures.

At the same time, industrial data are increasingly multimodal including time-series sensor streams, images/video, event logs, and maintenance text yet they remain noisy, partially labeled, and nonstationary. Large models offer a promising new paradigm: through pretraining on large-scale industrial data, they can learn more generalizable representations, adapt across scenarios, and support reasoning-oriented workflows for interpretable root-cause analysis. Coupling large models with digital twins further enables controllable virtual testbeds for fault injection, scenario expansion, and simulation-to-reality transfer.

Rationale and Objectives
Advanced equipment is now widely deployed in safety-critical and mission-critical settings such as intelligent transportation, coal-mine operations, and industrial manufacturing. FDD for these systems must move beyond component-level analytics toward:
o timely anomaly detection and fault isolation,
o degradation tracking and prognostics, and
o maintenance/operation decision-making under changing environments, scarce fault labels, and frequent software hardware updates.

This Research Topic aims to advance large-model-empowered FDD by leveraging foundation models trained on massive industrial sequences and multi-source observations to:
o learn transferable health representations,
o align heterogeneous evidence across modalities, and
o generate reasoning-capable diagnosis narratives with traceable support.

We further seek methods that incorporate system structure and physical constraints as priors, and that can be calibrated, validated, and deployed in realistic missions to measurably improve safety, availability, and lifecycle cost.

Scope (What this Research Topic covers)

This Research Topic welcomes contributions on:
1. Large time-series foundation models for fault detection, diagnosis, and prognostics under variable and nonstationary conditions
2. Multimodal large models that fuse sensor signals with images/video, event logs, and maintenance records for fault isolation and root-cause analysis
3. Large-model-enabled FDD for intelligent transportation systems and other safety-critical missions, emphasizing cross-scenario generalization and operational decision-making
4. Large-model-enabled diagnosis for engineering machinery (e.g., mining equipment) and industrial manufacturing equipment operating in harsh environments
5. Digital-twin-coupled large models, including simulation-based fault injection, scenario coverage expansion, and interpretable, deployable solutions

Article types and fees

This Research Topic accepts the following article types, unless otherwise specified in the Research Topic description:

• Brief Research Report
• Editorial
• FAIR² Data
• FAIR² DATA Direct Submission
• Hypothesis and Theory
• Methods
• Mini Review
• Opinion
• Original Research
• ... View all formats

Articles that are accepted for publication by our external editors following rigorous peer review incur a publishing fee charged to Authors, institutions, or funders.

Article types

This Research Topic accepts the following article types, unless otherwise specified in the Research Topic description:

• Brief Research Report
• Editorial
• FAIR² Data
• FAIR² DATA Direct Submission
• Hypothesis and Theory
• Methods
• Mini Review
• Opinion
• Original Research
• Perspective
• Review

Keywords: Fault Detection and Diagnosis, Multimodal Industrial Data Fusion, Digital Twin–Coupled Diagnosis, Trustworthy, Deployable Prognostics, Health Management (PHM)

Important note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.