Departures from basic geometries in architecture have historically often coincided with the development of new materials. This correlation is obvious in the 1950s, `60s, and `70s, during which developments in concrete and later plastics inspired architects and engineers to treat form in a less restrained manner.

The translation of a freeform into a built structure requires the development of new models of thinking from all project participants. It is essential that architects and engineers collaborate from the very beginning of a project. In the case of freeform architecture an important aspect of this collaboration is that the structural engineer has to “speak the language” of the architect and fully support the particular design approach.

Three approaches of a free form finding process.

1. The first approach (Shaping form) relies heavily on computational tools to generate the shape. This is a proper approach and all other ways are just simple simulations, as a non-alcohol beer. In the beginning of this approach architects have no preconceived formal idea. The “parametric design process starts with a briefing by the client. Architects translate program- and site-specific parameters into virtual forces using software environments such as Maya etc.  Architects set up a 3D matrix that was initially shaped according to the virtual forces. Adjacent objects on the site further impacted the shape through a series of specially designed force fields. The initial shape can be deformed and altered by parametric software, until the design parameters in architect’s option are sufficiently represented. The approximate shape can be also corrected for geometrical errors in order to establish 3D Master geometry of the project. This master geometry provides the dimensional reference for all project participants during design development and construction.
The design process is more or less defined as a linear increase in detailing of the architectural ides. Logically, designing a structure in such a planning process begins after the architectural design and is not intended formally to change the shape. From that point of view architectural design is not related to the structural logic of form, but structural design is nevertheless is a key factor in realizing the architectural idea.
The first approach we are given the external surface as the master geometry of the project. We generally have two options for developing the structure: we either design a system of linear or curvilinear structural members that support a secondary and nonstructural skin, or the skin itself is conceived as the primary load-bearing system and becomes “skin deep” – a surface –structure with shell-like behavior. Since the master geometry is fixed, we cannot optimize the structure through modifications of the overall shape. High local forces or bending moments have to be accepted: a structural optimization of the overall shape would call the underlying design approach into question.

2. The second approach illustrates a different approach. Here the design process does not start out in a virtual design environment; instead, the architects manually build a series of physical models, many of which are 3D digitized in order to correct and check the shape with respect to program and site in a computer-aided design environment.  The CAD-corrected data enables the building of more accurate physical model that often explore the implementation of the overall shaping strategy in partial and more detailed models of the project. This approach generally concentrates on the effects of the exterior surface and the interior spaces, and relies largely on physical model to verify that the original design intent is met. The most important difference from the first approach (form finding process) is that the second approach does not define a 3D master geometry as a dimensional reference before starting the design and structural development. Rather from finding turns out to be a kind of “iteration process” in which from changes are digitized and refined. In this approach we are given the database of the outer and inner surface and the interstitial space “between” results to integrate a structural system and embed all necessary mechanical elements. The loadbearing structure – often a series of steel frames – is hidden from user and architecturally almost irrelevant. In this “undercover role” the main part of structural engineering is a geometrical optimization that means to identify the layout of structural members in the interstitial space and optimize the arrangement so that it works in a structural way. The internal and external surface themselves act as enclosures without any primary load-bearing function; their geometry establishes boundary parameters for us that cannot generally be changed.

3. In the Third approach the conceptual design phase during the architectural competition does not rely heavily on computers, and physical models. The 3D model generated during design development is shaped to capture the design intent of the original handmade scheme. While following the initially proposed shape, it is built independently and does not contain digitized data taken directly from the physical models, like in the second approach. The third approach creates the digital model when translating the scheme from the conceptual status to a design development phase in a digital environment. Though the objective is to remain close to the original shape. And in contrast to the first approach the structural behavior could have influence on the shape, driven by the idea of possible structural systems and the understanding of their behavior. The external skin can be designed as a series of discrete layers , each responding to a specific set of functional requirements.