Context-aware Model Predictive Control framework for multi-zone buildings

Abstract

Today, Model predictive control (MPC) has largely been used to optimize energy consumption and maintain thermal comfort in buildings. However, to build an online MPC model in building, the dynamics of the physical system must be accurately modeled, which is a time-consuming and costly task. Neural network models help to overcome the modeling problems especially with the availability of historical data. This research presents a novel online data-driven control framework named Model Predictive Control via Genetic algorithm (MPC-GA) allowing the optimal operation of the heating, ventilation, and air conditioning system and has been experimentally validated in a multi-zone retail building. The MPC-GA combines an attention-based neural network time series multivariate prediction model with a MPC framework. The prediction model used a dual-stream neural networks based on multivariate time series of controlled and uncontrolled inputs. The attention mechanism is applied on controlled parameters to give them more weight to better predict the zone temperature. The prediction model is used as input for the optimization framework which minimizes: energy consumption, peak demand and discomfort during occupied hours under self-tuned setpoint, temperature ramp and equipment cycling constraints. A heuristic search algorithm using a genetic algorithm is used to solve the online data-driven MPC-GA models and obtain the future optimal combination settings of all controls for all the zones over a prediction horizon. The benchmark results showed that the MPC-GA outperforms RBC control systems with more than 50% and 80% reduction in energy consumption and discomfort respectively.