Adaptively learning surrogate model for predicting building system dynamics from simulation model

A building management system comprising:

• one or more processing circuits configured to:
• calibrate a simulation model for a building of interest;
• generate simulated data of system states using the calibrated simulation model;
• train a surrogate model based on the simulated data of system states from the calibrated simulation model;
• generate system state predictions using the surrogate model;
• re-train the surrogate model based on updated operational data;
• and
• generate an updated series of system state predictions using the re-trained surrogate model.

One or more non-transitory computer-readable storage media having instructions stored thereon that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• calibrating a simulation model;
• generating simulated data of system states using the calibrated simulation model;
• training a surrogate model based on the simulated data from the calibrated simulation model;
• generating system state predictions using the surrogate model;
• re-training the surrogate model based on updated operational data;
• and
• generating an updated series of system state predictions using the re-trained surrogate model.