论文部分内容阅读
用户出现识别被广泛应用于建筑环境中,例如包括需求控制的通风和安全工程。然而,目前大多数的传感器技术不能辨别出房间里具体的人数。为了解决这个问题,在卡内基梅隆大学的Robert L.Preger智能工作站(IW)内开发了一个复杂的环境传感器网络。结果显示,被测量的环境条件和用户状况之间有显著的相关性。它表明使用基于隐式马尔科夫模型的高斯混合模型在人数识别上能够实现83%的准确度。为了说明基于行为识别(例如人数和时间)带来的能耗影响,使用Energy Plus建立该智能工作站的模型,假定使用标准VAV系统,通过模拟计算来比较根据ASHRAE 90.1标准案例规定的作息模式和预测得到的两种模式之间的能源消耗结果。结果表明,在保证室内热舒适的同时,智能工作站可节约空调能耗18.5%。
User identification is widely used in the building environment, such as ventilation and safety engineering including demand control. However, most current sensor technologies can not tell the exact number of people in the room. To solve this problem, a complex environmental sensor network was developed at the Robert L. Preger Intelligent Workstation (IW) at Carnegie Mellon University. The results show that there is a significant correlation between the measured environmental conditions and the user’s condition. It shows that the Gaussian mixture model based on implicit Markov model can achieve 83% accuracy in the number recognition. To illustrate the impact of energy consumption based on behavioral identification (eg, number of people and time), a model of the smart workstation was built using Energy Plus, assuming a standard VAV system to simulate the rest and rest patterns and forecasts under the ASHRAE 90.1 standard case The resulting energy consumption results between the two modes. The results show that, while ensuring indoor thermal comfort, intelligent workstations can save air-conditioning energy consumption of 18.5%.