A model-assisted adaptive controller fine-tuning methodology for efficient energy use in buildings

Abstract

Building Energy Management Systems are finding widespread use for the holistic control of all energy-influencing elements of buildings and are responsible for ensuring an effective and parsimonious energy use. In most cases, fixed-logic controllers are deployed in the building to implement predetermined strategies. Good performance can not be guaranteed due to inherent uncertainties that can not be a priori ascertained, such as weather variations, occupant actions, and changes in the building state and characteristics. In this paper, a model-assisted tuning methodology is presented to adaptively and automatically fine-tune relevant controller parameters. In our approach, at the end of each day of the building operation, given “reasonable” predictions for the following day, and using an accurate thermal-simulation model to evaluate performance, a new set of controller parameters is generated to be used the following day. This way, good performance can be achieved using controllers with simple mathematical structure.