Interfacing between parametric associative and structural software

Computation and structural engineering have been going hand-in-hand for many years by structural analysis software and three-dimensional documentation software. However, software specifically designed for the design of structures during the early phases of the project is still scarce. The Structural Design Tools approach tries to deliver practical concepts and developments to make advanced computation more usable in structural design. Parametric associative design strategies are one of the base pillars of the approach as this technology provides a powerful tool for fast and easy modelling of complex structures. However, current parametric associative software is mainly focussed on providing tools for geometry, but tools to use the geometrical data in structural software is scarce and limited. This paper discusses a generic interface strategy between parametric associative software, GenerativeComponents (Aish 2005), and structural software. It also discusses the concepts and models behind the interface and proposals for possible future developments of the interface. manship of structural design and engineering can be achieved to come to more integral design. Opposed to other approach which follow a “single building model” approach, the Structural Design Tools approach follows a “multi-model” approach. This is based on the principle that the designing engineer uses many models during his process to gain insight and confidence in the behaviour of the structure and to gain control over the uncertainties and massive amounts of data and knowledge involved in the design process. The approach is therefore focussed on supporting freedom of interfacing, rather than restricting it, but encourages standardisation, such as IFCs (IAI 2005) as this shifts the custom development level towards the special cases. This approach is the successor of a framework model, called “openStrategy Form Finding” (Coenders 2004), which can contain form finding, structural optimisation, generative and iterative calculation techniques. Although the Structural Design Tools approach includes many concepts on different levels of abstraction and implementation regarding the concepts, tools and toolbox, this paper is mainly focussed on the combination of advanced geometrical structural designs, defined in the latest parametric associative systems (one of the base pillar of the approach), and interfacing to structural applications, such as structural analysis and three-dimensional documentation software. This last software is often also referred to as Building Information Models (BIM) (Bernstein 2004). 3 PARAMETRIC ASSOCIATIVE DESIGN STRATEGIES FOR STRUCTURAL DESIGN Parametric associative systems are systems, which in general generate an output, often geometry, from user-definable parameters and user-definable relationships (associations) between those parameters. This generation can be replayed every time one of the parameters or associations changes. This provides consistent design logic captured in rules in the form of parameters and associations. 3.1 Advantages of parametric associative systems The advantages of parametric associative design strategies for architectural, structural and other types of design are numerous. A few examples are: − Change is easy and fast. − Storage of design knowledge and logic in rules by parameters and associations. − Replayable and reusable design logic in the form of a generation process. 3.2 GenerativeComponents (GC) More than a decade ago, Robert Aish of Bentley Systems started design and implementation of a parametric and associative system, which supports many advanced modelling and programmatic concepts. The current name of this system is GenerativeComponents (GC) (Aish 2005). The system is currently in Beta stage (February 2007) and includes many advanced geometrical tools. Components or parametric associative objects in GenerativeComponents are called “features”. One of the important concepts of GenerativeComponents invented by Aish is called “replication”. 3.2.1 Replication Replication is a concept which basically implies that associations to every single object apply equally on any multiple of the same objects. For example, a line can be drawn to a single point, but also to a series or array of points with the effect that multiple lines will be created. Replication is a very powerful concept as it provides a very fast means of simply defining design logic in a single component and replicating that component over a large series or array of inputs, resulting in a multiple of those components. Multidimensional replication is possible, which supports large and complex arrangements of components. An often used example of replication is the population of components over a tessellated surface (see Fig. 1). Figure 1. Population (replication) of a simple bracing component (below) on a tessellated surface (above). 4 REAL-TIME INTERFACING TO STRUCTURAL APPLICATIONS In 2006 the first investigations and implementations were developed for a generic, real-time interface from GenerativeComponents to structural applications (Coenders 2006). At this point the interface supported simple elements and concepts, not fully taking advantage of advanced concepts, such as replication. For a complex structure every element had to be individually upgraded to its structural equivalent, leading to laborious input of data. 4.1 Interfacing to structural analysis software Using the interface it is possible to interface to structural analysis software, which can be used to analyse the structural performance of the model. The interface contains information for analysis, such as nodes, elements (1D) and loads. In the current implementation a plug-in interface to GSA (Oasys 2006) is included, but any other structural analysis software could be written to or interfaced to, given the restrictions in Paragraph 4.3. 4.2 Interfacing to 3d documentation software Although GenerativeComponents includes native support to create Bentley Structural elements for three-dimensional documentation or BIM modelling, the user might want to have the model available in other systems as well. In the current implementation a plug-in interface is available to Tekla Structures (Tekla 2006) which supports basic structural elements, such as steel and concrete beams. In the near future this will be expanded to include more advanced structural elements, such as concrete volumes and reinforcement lay-out. 4.3 Interfacing to other software Due to the generic set-up of the structural interface module, interfacing to other software than the software mentioned above only requires a new component which defines how to translate the managed information in the interface component to the application. If this application has a COM or .NET interface (Microsoft 2006), the link can be directly plugged into the interface by only providing how the data should be send across the COM or .NET interface. Otherwise it is also possible to use more lowlevel interfaces or write the information to files. Only when the application requires extra information from the managed data in the interface, the interface should be modified. 4.4 Concepts Several basic concepts have been used in the interface to form a structural representation of the building model and interface it to other applications. 4.4.1 Upgrading To transform the geometrical representation of the structure modelled in native GenerativeComponents features to a structural representation, a concept of upgrading has been used. By attaching a structural feature to its geometrical equivalent, the interface will be aware of the extra information required by the structural applications. In the near future also the new feature in GC of custom properties for geometrical objects will be investigated to achieve this goal, but it can be argued that upgrading makes it clearer for the user that a feature has a structural counterpart and where to edit the information added by this counterpart. 4.4.2 Support for replication As discussed above replication is a very powerful new concept in parametric associative design, which leads to great increase of efficiency and simplification of design logic. Therefore, interfacing the features which have been created by making use of replication is an absolute necessity to achieve an efficient and powerful interface. However, this poses problems as these features now can be a multiple (replicated features) of features under one name. Therefore, the interface needs to deal with these multiples of features as structural software requires a single collection of single entities. The interface deals with this problem by preprocessing the features and collecting and sorting each of the entity types. Each entity type gets its own numbering scheme and each entity gets assigned its own number to which can be referred by the interface. 4.4.3 One plug-in per application As projects, firms, institutes, etc. are not limited to the two or three software application above, other software should be interfaced to too. By providing a mechanism of extension and a modular set-up, the interface can be expanded by plug-ins for other applications. The interface layer has been set up in a modular manner so that the interface features which define the coupling between the interface data model and the applications to be interfaced to are placed in a separate feature. The user can simply attach his or her favourite interface feature to be able to interface to that application on the fly. 4.4.4 Pre-processing Pre-processing is required to deal with both replication and other operations which will need to take place to interface efficiently and create a consistent model before it is interfaced to the applications. To provide a modular, extendable base for preprocessing a rule-based processing system has been included in the interface. In the near future this system will be changed to plug-in features, so that the user can define his or h