National Technical University of Athens (NTUA)
Μαθηματική μοντελοποίηση της θερμικής άνεσης και της ποιότητας εσωτερικού αέρα και προχωρημένες τεχνικές βελτιστοποίησης του κτιριακού σχεδιασμού: εφαρμογές σε φυσικά αεριζόμενα κτίρια
Abstract
dc:descriptionThe present dissertation investigates the airflow pattern in naturally ventilated buildings using advanced computer simulation techniques such as field or Computational Fluid Dynamics (CFD) models. In particular, two-equation turbulence models, such as the Standard k-ε, the RNG k-ε, the “Realizable” k-ε, as well as the Reynolds Stress Model (RSM) are used to predict the flow field in and around the buildings considered. In the case of the Standard k-ε model the potential of improving the prediction accuracy is investigated by means of the following ways: (a) Application of a Two-layer Standard k-ε model, and (b) Modification according to the inlet flow variables profiles of the Atmospheric Boundary Layer (ABL). The aforementioned turbulence models and the proposed modifications are passive ventilation of a pilot building and of a real-scale building subject to actual passive ventilation (for which an experimental procedure is conducted). The CFD model is then used for the prediction of thermal comfort in naturally ventilated buildings by implementing the most appropriate thermal comfort indices (TCI), such as the Predicted Mean Vote (PMV), which, according to literature information, is modified for pure natural ventilation as follows: (a) Extension to account for the metabolic-rate reduction expected in warm environments and for the expectancy factor due to occupant’s habitat, and (b) Correction of the PMV for humidity effects, based on the Standard Effective Temperature (SET*). The present thermal comfort model is completed with the implementation of the Percentage Dissatisfied (PD) index to account for discomfort due to air draughts as well. Additionally, Indoor Air Quality (IAQ) indices are also implemented in the CFD model, i.e. the ventilation effectiveness related to the removal of common pollutants. Following the aforementioned coupled model (CFD-TC/IAQ) a database of input-output pairs is formed, which is used to train and validate radial basis functions (RBF) artificial neural networks (ANN) according to the fuzzy means method. The ANN distributions are then utilized to formulate a multiobjective optimization problem, which accounts for the available thermal comfort and indoor air quality indices, occupant activity level and special constraints provided by design guidelines. The problem is confronted using a gradient-based algorithm associated with a special handling of constraints and of the initialization space of the design variables (Initialization Grid Concept, IGC). Additionally, a technique to account for the non-dominated solutions of the objective functions according to the Pareto criteria is also introduced. Finally, for each case studied the optimal designs are presented by means of nomographs, which may serve as look-up tables for either decision-making strategies or automated systems to create “Intelligent Buildings”.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2010
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Stavrakakis, Georgios
- Σταυρακάκης, Γεώργιος
Subjects
dc:subject × 20- Υπολογιστική ρευστοδυναμική
- Θερμική άνεση
- Ποιότητα εσωτερικού αέρα
- Τυρβώδης ροή
- Φυσικός αερισμός
- Αρχιτεκτονικός σχεδιασμός
- Τεχνητά νευρωνικά δίκτυα
- Πολυκριτηριακή βελτιστοποίηση
- Computational fluid dynamics
- Thermal comfort
- Indoor air quality
- Turbulent flows
- Natural ventilation
- Architectural design
- Artificial neural networks
- Multi-objective optimization
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Χημικού Μηχανικού
- Engineering and Technology
- Chemical Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/25592