Applying artificial intelligence modeling to optimize green roof irrigation

• Green roof irrigation lacks cost-effective and reliable water conservation measures.
• We used artificial intelligence modeling to optimize green roof irrigation efficiency.
• Artificial neural network and fuzzy logic simulated changes in soil moisture.
• Real-time weather data trained the model to predict accurately soil moisture content.
• Our model maintained adequate soil moisture content whilst saving 20% water use.

Recent increase in green-roof installation has increased irrigation water consumption which could be wasteful using conventional watering management protocol. The knowledge gap in irrigation optimization to achieve water conservation could be filled. The complicated conventional approach uses weather and soil sensors to calculate watering needs, which is impractical and not cost-effective. This study employs artificial intelligence algorithms composed of artificial neural network and fuzzy logic, using weather data to simulate soil moisture changes to develop an optimal irrigation strategy. The artificial neural network is trained to predict soil moisture based on four daily weather variables: real-time air temperature, relative humidity, solar radiation, and wind speed. Fuzzy-neural network is applied to determine the irrigation time and watering volume. The simulation model successfully mimics the human brain in making irrigation decision. The artificial intelligence irrigation could maintain adequate soil moisture ranging from 0.13 to 0.22 m3/m3 and reduce 20% of water use with improved plant coverage. Since the evapotranspiration from living vegetation plays a key role in the passive cooling mechanism, better plant coverage could increase the thermal-energy performance of green roofs. The low-cost and effective technique can motivate the adoption of green roofs by alleviating the water consumption obstacle.