Search

An autonomous adaptive MPC architecture is presented for control of heating, ventilation and air condition (HVAC) systems to maintain indoor temperature while reducing energy use. Although equipment use and occupant changes with time, existing MPC methods are not capable of automatically relearning models and computing control decisions reliably for extended periods without intervention from a human expert. We seek to address this weakness. Two major features are embedded in the proposed architecture to enable autonomy: (i) a system identification algorithm from our prior work that periodically re-learns building dynamics and unmeasured internal heat loads from data without requiring re-tuning by experts. The estimated model is guaranteed to be stable and has desirable physical properties irrespective of the data; (ii) an MPC planner with a convex approximation of the original nonconvex problem. The planner uses a descent and convergent method, with the underlying optimization problem being feasible and convex. A year long simulation with a realistic plant shows that both of the features of the proposed architecture - periodic model and disturbance update and convexification of the planning problem - are essential to get the performance improvement over a commonly used baseline controller. Without these features, though MPC can outperform the baseline controller in certain situations, the benefits may not be substantial enough to warrant the investment in MPC., Comment: 12 pages, 7 figures

We propose a method that simultaneously identifies a linear time-invariant model of a building's temperature dynamics and a transformed version of the unmeasured disturbance affecting the building. Our method uses l1-regularization to encourage the identified disturbance to be approximately sparse, which is motivated by the slowly-varying nature of occupancy that determines the disturbance. The proposed method involves solving a convex optimization problem that guarantees the identified black-box model possesses known properties of the plant, especially input-output stability and positive DC gains. These features enable one to use the method as part of a self-learning control system in which the model of the building is updated periodically without requiring human intervention. Results from the application of the method on data from a simulated and real building are provided.