Investigating the thermal performance of green wall: Experimental analysis, deep learning model, and simulation studies in a humid climate

Abstract

Abstract Covering buildings with vegetation systems has been a significant feature of architectural design towards sustainability in recent years. This study reports an investigation on the green wall thermal performance compared to the bare wall on the northern facade of a 2-story residential building in the humid climate of Rasht during summertime. For experimental measurements, temperature and humidity data loggers were used for real-time data collection. Thereafter, an existing building was modeled in EnergyPlus for validation purposes. According to the results, a decrease in temperature and relative humidity was seen in the case of the building with a green wall. It was found that the green wall could drop the indoor temperature by 9% and also decrease the relative humidity level by 32%. Besides, in order to predict the green wall performance in a short time interval, a deep artificial neural network was trained from the experimental data and a 15-day weather dataset was collected and fed into the deep learning model. Moreover, the ENVI-met software is utilized to evaluate the effect of the green wall on the surrounding air. Findings indicated that the temperature in front of the green wall is slightly lower than the part of the wall without the plant. The highest temperate reduction was 0.36\xa0°C at 12 p.m., which is insignificant.