what is orbital
Orbital Overview
Introducing Orbital
Orbital is the first foundation model purpose-built for Refining and Petro-Chem. It combines three powerful models: a time series model, a physics-based model, and a language model trained on decades of chemical and process engineering expertise.
These models work in parallel, reinforcing each other, continuously retrieving and interpreting operational data in real time. Orbital returns insights in natural language that are fast, accurate, and grounded in real-world physics, process and chemical engineering.
Time Series Model
Connects to your DCS, historians, and LIMS to learn from raw, unsampled data. Allowing Orbital to forecast trends, detect anomalies, and adapt to drift and missing values.
Physics-Based Model
Extracts governing equations from manuals and literature using the LLM. Estimates unknowns from live data. Every prediction respects mass balance, energy conservation, and reaction kinetics — ensuring physical validity.
The Language Model
Trained on chemical engineering and petrochemical knowledge. It reads your P&IDs, SOPs, and work orders to retrieve context, explain reasoning, and recommend actions in clear, trusted process language.
Internal Components of Orbital
Once connected, data flows through Orbital’s inference stack.
This includes processing, authentication, temporal alignment, simulation, and large language model reasoning, all tuned for real-time decision-making.