Resources · Glossary
Industrial AI glossary
Plain-language definitions for the terms that matter on the plant floor and in the control room.
- Anomaly detection
- Live models that learn normal process behavior and flag deviations before they affect production.
- Closed-loop optimization
- Streaming approved predictions or setpoints back to control systems within defined safety boundaries.
- DCS
- Distributed Control System — the supervisory control layer in a process plant.
- Data Explorer
- Intelecy module for high-speed visualization and multivariable analysis of industrial time-series data.
- Forecasting
- Predicting future process values or outcomes so teams can act before thresholds are crossed.
- Historian
- System that stores time-series sensor and process data from plant equipment for analysis.
- Human-in-the-loop
- Governance pattern requiring an operator to approve AI-suggested actions before they are sent to control systems.
- Industrial AI
- Machine learning applied to industrial process and asset data to improve operational outcomes.
- Intelecy Gateway
- Secure data flow layer between industrial data sources and the Intelecy platform.
- MLOps
- Practices for deploying, monitoring, and maintaining machine learning models in production.
- No-code
- Tools that allow process experts to build, validate, and deploy models without writing code.
- OPC UA
- Open Platform Communications Unified Architecture — an industrial interoperability standard.
- Operating window
- The range of process conditions within which the unit produces on-spec, safe output.
- Prescriptive optimization
- Models that recommend setpoint moves or actions, with confidence and explanation.
- SCADA
- Supervisory Control and Data Acquisition — the system used to monitor and control industrial operations.
- Soft sensor
- Model that predicts a hard-to-measure value (e.g. lab quality) from related process tags.