Loading...

Optimizing a Vertical Farm Energy System using Intelligence Techniques

Asadi Shizari, Shaghayegh | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57826 (46)
  4. University: Sharif University of Technology
  5. Department: Energy Engineering
  6. Advisor(s): Roshandel, Ramin; Boroushaki, Mehrdad
  7. Abstract:
  8. The growth of urbanization and climate change, coupled with resource limitations, has intensified the demand for controlled environment agriculture (CEA) systems to ensure food security and environmental sustainability. Vertical farming (VF), as one of the most advanced CEA methods, offers significant advantages over traditional farming in limited urban spaces. However, the high energy consumption in these systems poses a serious challenge, leading to substantial economic and environmental costs. This study aims to develop a dynamic optimization model for vertical farming by leveraging machine learning models and advanced optimization techniques to minimize energy consumption and enhance production efficiency. To address the challenge of insufficient data availability, the required dataset was derived from an analytical model, which provided the necessary data for training. Using this dataset, various ML models were trained, and XGBoost was identified as the best-performing data-driven model. XGBoost accurately simulated the system's physics, achieving a high R² of 0.99, and providing precise results of the vertical farming process. Validation and sensitivity analysis were then conducted to evaluate the model’s accuracy in predicting various environmental conditions. Considering the critical influence of indoor air quality parameters- temperature, light intensity, carbon dioxide concentration and relative humidity- on system performance and energy consumption, the primary goal of the optimization was to fine-tune these parameters to minimize energy intensity (EI). The developed intelligent model was optimized using genetic algorithms, demonstrating high efficiency in providing an optimal operational plan for a 30-day cultivation period. This plan included detailed 30-day profiles for indoor air quality parameters. Notabley, during July, the hottest month of the year, the optimized profile reduced energy intensity from 11.43 kWh/kg (as reported by the Keyvan simulator model) to an optimal value of 8.84 kWh/kg, achieving a significant 22.7% reduction. This significant decrease in energy intensity, along with enhanced production efficiency, lowers economic costs and mitigates environmental impacts. These findings underscore the potential of optimized vertical farming systems as a sustainable and viable solution for urban agriculture
  9. Keywords:
  10. Machine Learning ; Vertical Farm System ; Extreme Gradient Boosting (XGBoost) ; Intelligent Optimization ; Indoor Quality ; Sustainable Agriculture ; Energy Intensity

 Digital Object List

 Bookmark

No TOC