This paper proposes an energy consumption diagnosis method and establishes an evaluation model based on data mining technology for evaluating the energy utilization of HVAC systems used in subway stations. First

through a correlation analysis

the key variables that influence the energy consumption of the subway’s HVAC system are determined

namely

the outdoor temperature and passenger flows. Then

according to the selected key variables

different energy utilization modes are generated using a decision tree

after which an energy benchmark is established on the basis of the energy consumption models. Finally

an energy consumption diagnosis method based on the energy benchmark of different energy utilization modes is applied to actual data of an HVAC system. The results of energy consumption diagnosis show that the energy levels can be influenced by the changes in the environment and passenger flow

but still conform to the energy benchmark. This energy consumption diagnosis method can diagnose the conditions of energy utilization and recognize modes of abnormal or low energy consumption

which can optimize energy utilization and provide a theoretical basis and practical reference for energy savings at a single subway station.