Fuzzy model based predictive control for active chilled beam systems

The paper proposes a model predictive control (MPC) strategy based on Takagi-Sugeno (T-S) fuzzy logic for Active Chilled Beam (ACB) terminal units in heating, ventilation, and air-conditioning (HVAC) systems. The ACB system is a multi-input-multi-output (MIMO) system with strong nonlinear characteristics. The relative humidity and the temperature are two outputs respectively controlled by varying the primary air volume flow rate and the chilled water inlet temperature. To solve the nonlinear and coupling issue, decentralized fuzzy modeling method is adopted to simplify the MIMO system into SISO control loops. By adopting the divide-and-conquer strategy for each individual loop, the complex nonlinear system is divided into a set of linear T-S models. In addition, a MPC strategy is developed in which different predictive controllers are designed for different T-S fuzzy rules and the overall controller output is obtained by weighted sum of the local controller outputs. The controller is validated in simulation, which is capable of maintaining the predetermined indoor relative humidity and temperature with predetermined constraints. Moreover, the proposed control scheme can greatly reduce computational complexity as the coupling effects from other loops are no longer required in the modeling designs.