Toward Greener Tertiary Buildings: Passive Parameter Optimization through Artificial Intelligence Forecasting and Statistical Analysis
TL;DR
Time-series forecasting and statistical screening techniques will be employed to determine the extent to which passive design affects the long-term HVAC energy demands, to show savings from 2021 to 2029.
Abstract
Cooling and heating account for a significant fraction of the overall energy consumed by tertiary buildings, especially in hot arid climates. In the current paper, time-series forecasting and statistical screening techniques will be employed to determine the extent to which passive design affects the long-term HVAC energy demands. The monthly heating and cooling demands of two buildings, each measuring 415 m 2 and used as administrative buildings, located in Errachidia, Morocco, will be simulated using Design Builder, serving as the basis for the analysis. The buildings will be designed for the period from 2016 to 2020. The first will be the baseline design, and the second will be the optimized design. The models employed will be the Seasonal ARIMA and the Long Short-Term Memory. The models will be evaluated using MSE, RMSE, MAE, and SI. At the same time, the analysis will be carried out using the ANOVA method to determine which passive design parameters affect the overall energy demands. The results will then be used to forecast the 2021-2029 period. The optimized passive design parameters, which will be used, will be the south-east orientation, 5cm air gap, and double glazing. The results will show that the savings from 2021 to 2029 will be 1,739,640 kWh and 893,246 kWh, which correspond to 181.2 kWh and 93.05 kWh, respectively, as the average savings each month for the 415 m2 building.