Pont de Bondy

Pont de Bondy is a somewhat “older” project from BIG that has been sleeping for a few years. The model I received had been made during the competition phase, coming with all the unsolved geometry problems that you’d expect; It was not following the concept drawings, wasn’t taking accounting for construction and wasn’t even following the structure. Just the regular good stuff.

THE ROOF

The first step was to get to know the project, get the actual structure planned for the ramp (mostly just axis) and the general intent of the design. The designers described the ramp to have 5 different steps:

• Constant slope straight (orange)

• Constant slope turning (pink)

• Varying slope turning (blue)

• No slope turning (green)

• No slope straight (yellow)

The problem might seem trivial, but ended up being slightly more fun to solve than expected. One major detail is that all the transitions would have to happen on radial axis (where columns are) that had been defined by the engineering team, since the transitions of each element meant also different built elements and sometimes different construction methods. Moreover, the solution had to fit under fixed heights and min / max slopes to fix the current design.

Few designs have been explored, here under the two main candidates. Black lines represents contour lines.
Note: on the dropped version, the outside edge goes up at a lower rate since it has a longer distance to achieve. The blue sections are the only place where slopes are changing from sloped to flat.

The dropped design (although exaggerated in the previous images) guarantied a better edge continuity between some steps. The selected design is much more in line with constructability, which was the decisional criteria.

For slightly more detail, in order to fit all the constraints, the design has been solved unrolled and involving a small optimization to find the best position for the circle (responsible for the blue part), visible on one of the images hereunder.

THE FACADE

Contrary to the renders above, and as you might have spotted in my images, the outside facade isn’t blind wood panels, but many thin vertical windows. Once again the current plans were not following the structure, and the sizes were all over the place, which is sort of expected in competitions. The design team also needed the window panels to rotate on a vertical axis to be at an angle from the edge of the roof.

The todo list:

• Find relevant vertical divisions while following the structures (columns), which means different widths

• Use construction details to locate the panels in 3d

• Define and model the rotations

• Build excel file to track and compare designs

• Push to Revit

After finding the mullion divisions so that columns always meet with one mullion and located all the elements, I worked on the tilting of each panels relative to the curved facade.
Many parameters were available, but in the end the panel transition would happen in one columns interval, twisting from 0 to +12 degrees.

Some more images hereunder illustrating some of the differences between the panels.

ETFE ROOF

The EFTE roof was supposed to be a very quick and dirty yet simple modeling to have some visuals of it in the Revit file for administration documentations. Details were rough, the deadline was short and the main goal was for inaccurate views and sections only. I therefore decided to stay off of kangaroo and and find a geometric approach that would look somewhat like layers of inflated elements.

All the elements were sent to Revit as a generic model. All the geometry (7 different layers in total) were all non intersecting solids. All the pipe geometries required a special attention (and modification) in order to be rendered properly by Revit by default.

Antoine - 2024