Cluster 1: Bike frame patterns

Comparison of 2D patterns from interconnected bike frames

This study expands on previous work with an analysis of different variations of patterns and their design flexibility. Depending on the joints used, fames are oriented differently, and as a result, some patterns are more flat (right), or have a larger depth/height because adjacent frames meet in a 90° angle (middle, and left). Relating to the number of frames comprising a cell, patterns are more or less capable of adapting to geometry variation in the set of parts. With higher flexibility for adaptation, however, assemblies also become softer and therefore less stable/stiff as a construction.

Simulations of global form adaptation to minimize deviation angles at co-linear connections show the varying capacity of patterns to integrate differing geometries. The numbers at the bottom corner display the average deviation and the local maximum, also highlighted with a pink circle.

Pattern 1
Pattern 2
Pattern 3

Besides pattern logic, the constellation of parts also matters for how well geometric variation is compensated. 1000 random constellations of an identical set of parts are tested for each pattern, the statistic at the bottom right shows the distribution of max. angles across all tested variations. The study shows, that the three patterns produce different statistics, indicating the respective adaptability. Pattern2 performs best in this regard, pattern2 second best, and 3 least – which aligns with the number of parts required to form a cell (see above).

Pattern1 – max. angles between 2.07° – 18.4°
Pattern2 – max. angles between 0.88° – 5.27°
Pattern3 – max. angles between 3.43° – 20.07°