Distributed Model Predictive Control of a HVAC System via a Projected Subgradient Method

摘要

This paper presents a nonlinear model for the climate control of temperature, relative humidity and CO_2 concentration in a multi-zone HVAC system. A constrained optimization problem has been formulated to meet the objectives of good thermal comfort and energy optimization. Soft constraints have been introduced to increase the feasible region and provide balance between these two objectives. A distributed model predictive control (MPC) algorithm based on projected subgradient method is proposed to design the controller for the HVAC system. The effectiveness of the proposed algorithm is verified by comparing with the existing works through numerical simulation.