Ambient temperature and solar radiation prediction for predictive control of HVAC systems and a methodology for optimal building heating dynamic operation

Abstract

An optimal predictive control methodology has been developed for real-time dynamic operation of building heating systems. It integrates a weather predictor, a setpoint optimizer, a feedforward control scheme, and an adaptive generalized predictive control algorithm in order to achieve high building thermal performance. A new weather predictor simplified through normalization makes it feasible to quantify the qualitative weather forecast for solar radiation. A heating process model is estimated by means of recursive least-squares techniques with a U-D factorization algorithm and several supervision rules. Using the predictions of solar radiation and ambient temperature, and the identified model of the heating process, the zone setpoint is optimized through on-line simulation. The adaptive generalized predictive controller (GPC) associated with a feedforward control scheme has been improved with a new algorithm. This new control algorithm is capable of compensating for large thermal lags and dynamics in the feedback control loop. Results from the weather predictor are presented.