AI-Powered Root Cause Identification for Industrial Incidents

Department

Development department — PFE 2026–2027

Context

SIOTH® Root Cause Identification (RCI) helps operators trace industrial incidents back to their causes using configurable cause-effect models.

Project

Bring artificial intelligence into RCI by learning from historical data to identify the most likely root causes, anticipate incidents before they occur and continuously improve diagnosis accuracy.

Platform capabilities • Root-cause ranking with confidence scores. • Incident prediction from event and sensor trends. • New cause-effect links mined from historical data. • Explainable causal paths with supporting evidence. • Operator feedback to retrain the models.

Main topics • Study the RCI graph model and causal AI techniques. • Build an incident dataset from RCI and historian data. • Develop root-cause ranking and prediction models. • Integrate AI services into RCI using a .NET API and Angular UI. • Validate the solution on historical and simulated incidents.

Deliverables • AI services and RCI integration source code. • Angular views for ranked causes and predictions. • Documentation including report, design report and test reports.

Expected results

Faster diagnosis with ranked and explained root causes, early warning before incidents escalate, and a causal model that improves with every incident.

Keywords

Root Cause Analysis, Causal Graph, Bayesian Networks, Machine Learning, Predictive Analytics, Time Series, Rule Engine, IIoT, .NET/C#, Angular, SIOTH.

Development department — PFE 2026–2027

Context

SIOTH® Root Cause Identification (RCI) helps operators trace industrial incidents back to their causes using configurable cause-effect models.

Bring artificial intelligence into RCI by learning from historical data to identify the most likely root causes, anticipate incidents before they occur and continuously improve diagnosis accuracy.

Platform capabilities • Root-cause ranking with confidence scores. • Incident prediction from event and sensor trends. • New cause-effect links mined from historical data. • Explainable causal paths with supporting evidence. • Operator feedback to retrain the models.

Main topics • Study the RCI graph model and causal AI techniques. • Build an incident dataset from RCI and historian data. • Develop root-cause ranking and prediction models. • Integrate AI services into RCI using a .NET API and Angular UI. • Validate the solution on historical and simulated incidents.

Deliverables • AI services and RCI integration source code. • Angular views for ranked causes and predictions. • Documentation including report, design report and test reports.

Expected results

Faster diagnosis with ranked and explained root causes, early warning before incidents escalate, and a causal model that improves with every incident.

Root Cause Analysis, Causal Graph, Bayesian Networks, Machine Learning, Predictive Analytics, Time Series, Rule Engine, IIoT, .NET/C#, Angular, SIOTH.