Receding-Horizon Chiller Operation Planning via Collaborative Neurodynamic Optimization
Zhongying Chen, Jun Wang, Qing‐Long Han
Abstract
Optimal chiller loading is crucial to reduce energy consumption in chiller operation planning. In existing methods for planning with heterogeneous chillers, minimum-up/down-time constraints are not imposed. This paper addresses receding-horizon chiller operation planning via collaborative neurodynamic optimization. A mixed-integer optimization problem with minimum-up/down-time constraints is formulated for receding-horizon chiller loading with heterogeneous chillers. It is then decomposed into a binary optimization subproblem and a global optimization subproblem, to facilitate the planning process. A neurodynamics-driven algorithm is proposed based on paired discrete Hopfield networks and projection neural networks to solve the subproblems alternatingly and iteratively. Experimental results based on the specifications of two chiller systems are elaborated to substantiate the efficacy of the proposed method.