Module 6 - Evaluate Your Alternatives
Overview
For this project, I created a tower composed of three main tubes. Each tube rotates around the tower’s centre axis to create this torsion effect. The tubes all have different lengths, which gives the tower some depth and interesting geometry. The user can adjust and flex the tower height, its torsion, the tube thickness at three points, as well as the length offset between each tube end.
Figure 1: Parametric tower, personal design
Figure 2: Parametric tower with adjusted panel size and colour
The goal of the assignment was to find the tower parameters that would be best to provide natural ventilation in the building. For that, we flexed three different metrics:
1) The Daylight Factor looked at the tower's potential for receiving natural light. For that, we looked at the percentage of floor area within 6 feet from the outer shell of the tower. The bigger the percentage, the more daylight the building can receive.
2) The Height to Volume Ratio computed the tower height to its overall volume. A larger ratio means that the tower is higher than it is thick/wide. This would help improve the potential for natural ventilation by creating a greater temperature gradient inside the building, and therefore, improving natural ventilation inside the building.
3) Finally, the Cumulative Solar Insolation looked at how much solar energy the building can receive on its outer surfaces. The greater the value, the more heat can be generated from solar energy, thus helping create a larger temperature gradient inside the building without the help of HVAC systems.
Results
The following section shows the results I obtained for this assignment. Table 1 below displays the building performance of the tower when its height and middle section radius are flexed.
Table 1: Building performances for the rounded triangle design
From this analysis, we can see that we achieved our desired design to optimise natural ventilation for a 750-ft tower, with a middle section amplification factor (MidAmp) of 4.5. The estimated construction cost for this design is $1.45B. Our second and third best choices are for towers 700-ft tall and with a MidAmp of 4.5 and 750-ft tall and with a MidAmp of 4.75 respectively. The estimated costs for both designs are $1.32B and $1.55B respectively.
Ideally, we would like to minimise our project cost. First, we estimated the construction cost per square foot of floor area. This cost “will grow linearly from $500 per SF at the ground level to $1000 per SF at 750’ above the ground”. Then we could compute the overall construction cost from the estimated gross floor area.
Our results indicated that a 550-ft tall tower with a 4.5 MidAmp seems like the most economically sound option with an estimated construction cost of $986M. On the other hand, a 750-ft tall tower with a 5.5 MidAmp would cost close to $1.85B.
Recommendations
For this project, we would recommend choosing a design aiming at improving natural ventilation, while keeping an eye on the overall project cost. The 750-ft tower with 4.5 MidAmp yields the best scaling factor, but its cost is still considerable. Therefore, if the project team looks mostly at improving natural ventilation, we would recommend going with this option. However, by choosing a tower 50 feet shorter, we could cut down the cost by a whopping $110M, while keeping a very reasonable natural ventilation potential.
Modelling approach
The following steps describe my modelling approach for this assignment. For this part of the project, we used created our own design in Dynamo, as shown in figure 1 in the Overview section. We then flexed this form to our liking, and gather some expansive results in order to complete our building analysis.
Step 1 - Creating the structure [old]
Gathering user input
Firstly, we need to gather some input from the user in order to control the shape of the structure. The inputs chosen for this structure are shown in figure 3 below.
Figure 3: Prompting user inputs
Our structure is shaped on circles placed at the extremity of three equilateral triangles. These triangles are centred along the same axis but have different orientations, sizes, and heights based on the user inputs. The triangular base slider helps control the overall size of the triangle. The section height slider helps control the height of the selected section (either the middle or the top section). The radius amplificator slider allows to augment or diminish the size of the circles place on the triangle. When closer to 0, it gives the tower a thin shape, and vice-versa when closer to 2. Finally, the section rotation slider helps the user to choose by how many degrees a section is oriented. When selecting 60°, the section will be rotated 60° clockwise on the z-axis.
Create the wire structure
After gathering the user inputs, the wire structure for our tower can be created. The different steps are shown in figure 4 below.
Figure 4: Defining our geometry
As mentioned in the previous sub-section, we created three different equilateral triangles at three different heights. We then placed nine circles at each end of the triangles. We then rotated the circles along the z-axis according to the user input in order to create the tower’s twisting shape. We then selected each circle based on its height.
Merging geometries and lofting our structure
After having isolated the circles based on their height, we joined them using the List Create node. We matched a circle from the bottom section, with its corresponding circles in the middle and top sections as shown in the Merge Geometries group from figure 5 below. We then lofted and created a solid from the structure in order to obtain our final shape.
Figure 5: Merging wire structure and lofting the final structure
Step 2 - Measures of interest for our analysis
Extracting our measures of interest [old]
From the solid described in step 1.3. before, we could then extract some valuable measures for our building analysis.
Figure 6: Extracting the measures of interest from our structure
First, using the Solid.Volume node we could compute the volume of the tower. We then extracted the exterior surface area as shown in the group Extract Gross Surface Area. We exploded our solid into several surfaces. We then took the normal vectors of each surface in order to select the ones we wanted for our analysis. Z-vectors of -1 and 1 were excluded as they represented the roofs and floor of the structure. We then added the remaining surfaces to obtain our final gross surface area. Finally, we extracted the floor areas of our building. As shown in the Evaluate and Extract Floor Surface Area group, we created rectangular planes at every floor level. Each plane would “cut” our solid on the horizontal plane. We would then store and add each resulting surface to obtain our total floor surface for the building.
New measures of interest
The new measures of interest are described in the Introduction of this assignment.
The Daylight Factor was determined using the floor surface of the building. In our custom node, each floor surface outside the perimeter is offset by 6 feet to the inside. This new curve helps create a new smaller surface. By subtracting the actual floor surface with the newt offset, we can obtain the floor area affected by daylight. Then, we can simply take that value and divide it by the overall floor surface to obtain the percentage of floor area that receives daylight. The logic for this metric is shown in figure 7 below. Additionally, the Height to Volume Ratio was computed by diving the tower height by its volume.
Figure 7: Evaluate the percentage of daylight received
The Cumulative Solar Insolation was determined using a custom node. By setting the project location, time of the day, and time of the year, we were able to determine the heat energy from the sun on the exterior surface of the building. The logic is shown in figure 8 below.
Figure 8: Evaluate the cumulative solar insolation
Custom node for the structure
Figure 9: Custom node inputs and outputs
Step 3 - Analysing our results
Flexing two parameters
For this example, we tested two variables: the tower height and the MidAmp. As shown in figure 10 below, we created a cartesian product for both variable ranges in order to compute our different scenarios.
Figure 10: Testing two alternatives
As shown in figure 11 in the next subsection, we used the Function.Apply node to test those two parameters within our new custom node.
Creating a scaling factor
After testing our different results, we tried to create a scaling rank to choose the best design option for our example. As shown in figure 11 below, we took the maximum and minimum of each list to weight our scaling factor.
Figure 11: Gathering results and taking min and max from each list
Then, we chose three different scaling factors for the three metrics we wanted to test. We plugged those factors as well as the lists maximums and minimus in a custom node in order to generate our ranking scale for each case. We exported the results to an excel document where we could later analyse the information and select our best design choices. The logic is shown in figure 12 below.
Figure 12: Creating a ranking scale for our analysis
Providing visual feedback for our best result
In order to provide visual feedback to our analysis, we needed to first select the surfaces to panelise. As shown in figure 13 below, we were able to isolate all the exterior surfaces of the building.
Figure 13: Selecting specific surfaces
From there, we extracted these surfaces again for our best-case scenario, and panelise them. We choose a specific type of panel from which we modified its opening and colour to match the sun directness on the surface. This logic is show in the figure below.
Figure 14: Providing visual feedback for our best-case design
The results are shown in figure 2 in the Introduction section.