Advanced modelling of green roof-enabled Plus-ZEB energy systems: A synergistic approach using Taguchi design and neural networks

This study develops a comprehensive Energy-Plus model of a single-story residential building in Tehran's climate, with a 200 m2 footprint and roof area (20 m × 10 m), to assess the impact of green roofs on heating and cooling demands. The base case features a conventional insulated roof comprising a 100 mm heavyweight concrete layer, ceiling air space resistance, and a 19 mm acoustic tile layer. The green roof configuration is examined as a pathway toward Plus-Zero Energy Building (Plus-ZEB) performance. Using the Taguchi L32 (2 × 43) method, the effects of four parameters, including stomatal resistance, plant height, substrate thickness, and soil surface roughness, are evaluated under diverse scenarios. A neural network (NN) model is then trained on simulation data to predict energy consumption for other parameter configurations within the studied ranges. Results show that a fully vegetated roof can reduce annual heating demand by 9.09 % (5.0 MWh/yr) and cooling demand by 10.08 % (2.6 MWh/yr) relative to the baseline. Among the examined parameters, substrate thickness had the most significant influence on energy savings, followed by plant height. These findings guide early-stage design, highlighting optimized green roofs as an effective strategy to cut residential energy use and support sustainability.