Whole building operation optimal control system based on an occupancy sensor network

Building electromechanical facilities yield more and more significant power consumption. This motivates researchers to work on how to improve building energy efficiency. However, it is difficult to improve building energy efficiency at the scale of whole building currently. Part of the difficulty is the lack of building energy management system, but an even greater deficiency arises from absence of optimal algorithms appropriate to all facilities in whole building. To address this problem, a integrated optimal control method has been proposed in this paper. The integrated optimal control system consists of a building facility internet of things, an occupancy sensor network and adaptive dynamic programming algorithm. To avoid shortcomings in available sensor technology, cameras, passive infra-red sensors, RFID sensors and CO₂ concentration sensors have been deployed to set up a hybrid occupancy detection sensor network. Meanwhile, a information fusion model based on sensor-mapping spiking neural networks has been put forward as a data analysis tool for determining when and where people occupy a given space. It is very critical to control and optimization for complex systems such as the system constituted by all electromechanical systems in a building. A novel adaptive dynamic programming method has been described in this paper. It learns from environment of a building and generates a series of optimal control strategies to preserve human comfort and improve energy efficiency.