An artificial intelligence-based system for implementing proxy models for physics-based simulators.

Abstract

The simulation method includes using input data from a plurality of different sources of operational data (215) including time series data to provide physics-based simulation models (220a, 220b) including model parameters (220a1, 220a2) for simulating a physical process, the physics-based simulation models generating output data including simulated forecasts calculated using the model parameters, and an artificial intelligence (AI)-based system (210) including an AI-based proxy model (210a) that, in response to receiving updates to the input data, processes the updated input data to generate a proxy forecast (211) for at least one selected forecast from the simulated forecasts or variables derived from the simulated forecasts as a substitution or supplement for the selected forecast or variables derived from the selected forecast (102).

Images