Navigating Complex Models in Collaborative Work for (Sustainably) Integrated Design

Increasingly intensive use of computational techniques such as parametric-associative modeling, algorithmic design, performance simulations and generative design in architecture, engineering and construction are leading to increasingly large and complex 3D building models which in turn require increasingly powerful techniques in order to be manipulated and interpreted effectively. Further complexities are of course due also to the multi-disciplinary nature of building projects, in which there can be significant variation and even conflict among the aims of architects, engineers and builders, as well as owners, occupants and other stakeholders in the process. Effective use of model information depends to a large extent on sense-making, which can in some ways be helped but also hindered by schemes for organizing the information contained. Common techniques such as layering, labeling (aka tagging) and assignment of various other attributes to model objects have significant limitations especially those arising from general problems of language, ontology and standardization, as well as but distinct from issues of interoperability both with respect to locating the desired items in a 3D building model and also with respect to displaying the objects in informative ways which effectively assist collaborative design and decision-making. Sustainable design in particular is an area generally requiring a high level of inter-disciplinary collaboration to achieve highly integrated designs which make multiple use of the elements and systems incorporated (though integrated design may also be pursued without explicit aims of sustainability). The proposed paper describes ongoing research concerning alternatives to the currently common techniques for locating and displaying information in 3D building models in support of sense-making to promote collaborative and integrated design. These alternatives comprise on the one hand interactive geometric-content-based methods for search and classification of model objects as an alternative or complement to common assigned-attribute-based methods and on the other hand visual analytic techniques in contrast to existing, relatively static tabular and "physical" views which can help to increase the informativeness of the geometric data within the model, as well as the non-geometric data that is attached to geometric objects (e.g. as in the cases of BIM and various types of CAE performance simulations). Tests undertaken with architects and engineers in practice and academia to evaluate the proposed methods are also described. Finally conclusions are drawn regarding these methods positive present performance and some of their shortcomings, as well as indicating directions for future research concerning the methods refinement and extension to help 3D building models become more effective components of the design process than they are at present, both with respect to these models present levels of complexity and especially with respect to their anticipated increasing complexity in future.