Decentralized optimization for energy savings for a HVAC chiller plant

Improved control strategies of building HVAC (heating, ventilation, and air-conditioning) systems can significantly improve energy savings while maintaining human comfort requirements. This paper presents an optimization problem for the setpoints of the control variables to give the lowest chiller plant energy consumption. The optimization problem is nonlinear, strongly coupled, and has both discrete and continuous control variables. A decentralized optimization method is given that divides the system into multiple subsystems and then solves the subproblems based on local information. Numerical results demonstrate that even with limited knowledge of neighboring subsystems, each subsystem can find the optimal solution of the original optimization problem. This method has broad applications since it does not require global system knowledge.