Energy consumption modeling by machine learning from daily activity metering in a hospital

摘要: Hospitals are large buildings that consume a great amount of energy mostly due to their continuous consumption needs, high consumer medical equipment, and special requirements thermal air conditions. Reliable dynamic simulation is chimera because the complex design behavior these buildings. Therefore, monitoring-based methods arise as plausible alternative. Its main drawback, however, lack enough data generate statistically robust models. The paper faces this problem thanks helpful contribution collaborative hospital which was able daily electrical for period six years. Besides, thirteen variables summarize activity also included. results show how machine learning techniques models accurately predict based on weather conditions measurements. obtained useful more specific saving strategies, efficient economic investment retrofitting existing better management cost in large-scale