Uncertainty method and sensitivity analysis for assessment of energy consumption of underground metro station

An accurate calculation of energy consumption is a precondition for energy underground metro station design, which determines the amount of energy geostructure. This study develops a simplified deterministic method for the underground metro station energy performance that accounts for weather and interior heat gain uncertainty. A Monte Carlo technique with Latin hypercube sampling is then employed to confirm the probability distributions of the peak load, average yearly load and annual energy demand, and compared to deterministic method to improve the design robustness. The sensitivities of 14 input variables with respect to the underground metro station energy performance are discussed through three sensitivity methods. The simplified deterministic method is more accurate than the DeST and American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) in describing the fluctuation of the underground metro station energy performance. The uncertainty distribution of energy performance is advantageous to the system design, considering both safety and economy. Moreover, a comparison with the deterministic method is performed to determine the reasonability of the safety factor 1.2, which is usually used in practical programs. The peak load is dominated by outdoor parameters, while there are no control parameters for the average yearly load or annual energy demand.

Automated summary

An accurate calculation of energy consumption is crucial for the design of underground metro stations. This study presents a simplified deterministic method that considers weather and heat gain uncertainty to evaluate the energy performance of metro stations. Monte Carlo technique was employed to confirm the probability distributions of load and energy demand and improve design robustness. Sensitivity analysis of input variables was conducted, and the simplified deterministic method proved to be more accurate in describing energy performance fluctuations compared to existing methods.