Completing previous methodology steps, it is now time to actually try and address the issue. This section will explain the approach followed in order to find the optimum solution. After careful consideration and deep study of windbreak literature and research, it was decided to follow a 2-step procedure. Initial step was to gather data from the simulation environment and try to extract information of which exact areas are the most problematic with regards to leaving large quantity of wind entering the passage between the buildings. This amount of wind once found trapped between the buildings, had no way out causing the phenomenon under investigation. So, the approach followed here had the objective of trying to avoid wind being trapped between the buildings. Gathered data clearly showed that significant amount of wind trapped was entering through the areas shown in below diagram (Fig_35).
Following step was to implement Genetic Algorithms in order to find the windbreak characteristics needed to resolve the issue. GA was configured to try to find the optimal solution with regards to finding the most porous windbreak structure (that is to be implemented by tensegrities) addressing the phenomenon. This was chosen as tensegrities are porous structure from their nature and it could prove very hard or even impossible to try and use them to construct a windbreak with very low permeability percentage. Another reason for choosing this optimization tactic is that porous structures would be very easy and cost effective to construct and also are less interventional into the environment as they will only affect the specific issue and leave things like temperature and sunshine uninfluenced.
Under this scope, GA were carefully designed and used in order to act as an optimization technique on this effort. The idea was to create different amount of blockages (solid internal boundaries) and always construct configurations, which will act as the windbreaks. In GA terms, a whole configuration of is considered as an individual with genes consisting of the amount of blockages and their location. Extra care has been taken here to always make sure that these blockages do not randomly appear out of nowhere but are always connected in a way that they could define a structure that can be built. This was done by applying constraints and running extra checks during the genes generation to always make sure that every individual is using a valid configuration.
Following step was to implement Genetic Algorithms in order to find the windbreak characteristics needed to resolve the issue. GA was configured to try to find the optimal solution with regards to finding the most porous windbreak structure (that is to be implemented by tensegrities) addressing the phenomenon. This was chosen as tensegrities are porous structure from their nature and it could prove very hard or even impossible to try and use them to construct a windbreak with very low permeability percentage. Another reason for choosing this optimization tactic is that porous structures would be very easy and cost effective to construct and also are less interventional into the environment as they will only affect the specific issue and leave things like temperature and sunshine uninfluenced.
Under this scope, GA were carefully designed and used in order to act as an optimization technique on this effort. The idea was to create different amount of blockages (solid internal boundaries) and always construct configurations, which will act as the windbreaks. In GA terms, a whole configuration of is considered as an individual with genes consisting of the amount of blockages and their location. Extra care has been taken here to always make sure that these blockages do not randomly appear out of nowhere but are always connected in a way that they could define a structure that can be built. This was done by applying constraints and running extra checks during the genes generation to always make sure that every individual is using a valid configuration.
Fig_35