The invention discloses a hospital
electricity consumption real-time prediction method and
system based on working condition domain division, and the method comprises the steps: extracting the geometric and topological relation information of a
hospital based on a BIM model of a hospital building, and determining the deployment number and positions of
Internet of Things sensors; extracting hospital
electricity consumption historical data through
the Internet of Things sensor, obtaining hospital
electricity consumption data and electricity consumption influence parameters in real time, and screening input variables; dividing a plurality of working condition domains in combination with hospital electricity consumption historical data; extracting high-level features of the
energy consumption diagrams of the working condition domains, extracting statistical features and
time sequence features from the real-time
power consumption influence parameters, and fusing the statistical features and the
time sequence features to generate multi-mode feature vectors of the working condition domains; training a hospital electricity consumption prediction model of each working condition domain; and after the training is completed, inputting hospital
power consumption influence parameters obtained in real time, and outputting hospital
power consumption predicted in real time. According to the method, multi-
source data are combined, working condition changes are dynamically adapted, and high-precision real-time prediction is achieved.