Engineering with Computers 7, 1-9 (1991)
Ewngineering Computers 9 Springer-Verlag New York Inc. 1991
A Feature-Based Approach to Structural Design M.K. Zamanian 1, S.J. Fenves l, and C.R. Thewalt 1, and S. Finger 2 Civil Engineering Department ~and Robotics Institute 2, Carnegie Mellon Umversity, Pittsburgh, PA, USA
Abstract. Despite the continuing improvements in computeraided design (CAD) systems and improvements in geometric modeling, most CAD systems are used as advanced drafting and drawing management tools by structural designers. A computer model of a structural design usually is generated by creating a detailed geometric model of the primitive components of the design and then attaching attributes, such as physical properties and loading conditions, to the various geometric components to reveal the structural characteristics of those components. This bottom-up approach has been inherited from early drafting techniques and contrasts sharply with a structural designer's natural way of thinking and reasoning about the design. Geometric features, on the other hand, provide high-level abstractions of design information and can be tailored to a designer's specific engineering needs. In this paper the advantages of using feature-based techniques in structural CAD systems are discussed. These techniques provide better modeling primitives for users and superior data models for CAD systems for reasoning about the geometry, topology, and engineering properties of a structure.
1
Introduction
Most engineered artifacts designed or analyzed using CAD systems are complex. The geometric models that represent these artifacts are often produced by many people over many months and are based on large amounts of data and knowledge about various aspects of engineering. Prior to the use of CAD technology, the only means of recording and transferring this information were engineering drawings and reports. As a result, the early CAD systems simply provided computer tools that enabled designers to generate and modify drawings and reports faster and more efficiently. Furthermore, these CAD systems introduced canonical representations of geometry in wireframe form that
Offprint requests: M.K. Zamanian, Civil Engineering Department, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
facilitated the storage and exchange of geometric information in digital form. The wireframe representation of geometry appeared suitable for twodimensional drawings but had serious deficiencies, such as ambiguity and invalidity, in modeling of three-dimensional objects. These deficiencies eventually lead to the development of solid modeling techniques, such as constructive solid geometry (CSG) and boundary representation (B-Rep), that provide a mathematically sound basis for representing three-dimensional objects. Most of the recent CAD systems use solid modelers as their primary geometric modeling component. The use of geometric modeling, computer graphics, data base management, and other computer technologies has increased the efficiency of engineering design offices in terms of production, management, and communication of their complex drawings; however, these techniques have provided little assistance to the designers at the early stages of the design process. On most of the current CAD systems, one can automatically compute mass properties or generate a finite element mesh for a three-dimensional computer model, but the designer is still responsible for creating the initial (often detailed) design on a sketch-pad before it can be modeled on a CAD system. Furthermore, the modeling capabilities of the current CAD systems are centered around the geometric modelers and consequently force users to spend much time creating and modifying detailed geometric information. Only after the geometric model is complete can engineering attributes, such as physical properties, be added to the model. This bottom-up approach is contrary to the thinking process of most designers and is due primarily to the fact that CAD systems were created as drafting aids without much concern for the designers' needs. Several groups of researchers in the fields of information modeling and artificial intelligence [1-3] have suggested better methods for representing dif-
2
M.K. Zamanian et al.
ferent aspects of the design information in CAD systems. A result of this effort has been the development of intelligent CAD systems that can reason about designs [3-5]. The intelligence of a CAD sys 7 tern can be measured by the system's ability to understand higher-level concepts and to execute tasks defined in terms of these concepts [3]. However, purely geometric representations using the available solid modeling programs are unable to provide the information necessary for reasoning about the nongeometric aspects of design. Research in the area of features has resulted in many promising techniques for combining engineering data and knowledge with geometric information. Almost all the research on features has been in the domain of mechanical design, largely because the primary goal of the mechanical CAD systems has been to provide concise, accurate representations of mechanical parts along with their corresponding manufacturing processes. Furthermore, the need for integrated design environments and automated information processing techniques for complex mechanical designs has prompted extensive research for better representation of geometry and topology of both completed and in-progress designed objects [4]. During the past few years a growing number of researchers has investigated similar integration and automation issues for architectural and structural design problems [6-8], However, most of the current research has concentrated on issues such as data base management, expert systems, and graphic user interfaces without much emphasis on highlevel, geometric modeling primitives for representation of structural designs. In fact, wireframe and polygonal schemes are still widely used for geometric modeling of structural designs. In this paper we propose to use features for structural design. Features are presented as high-level modeling primitives that facilitate the generation and modification of structural designs in a CAD environment. Moreover, the information contained in these features is essential to many application programs for reasoning about the various characteristics of a structure.
2
Features
Features are abstract entities that combine functionally related elements of a model [9]. During the past several years many researchers have proposed using features as a natural form of communication among designers, analysts, and manufacturers about the topology and geometry of designed arti-
facts [10-14]. Features were first introduced for combining machining information with solid models of mechanical parts. Perhaps the most commonly known type of features used for this purpose are form feature. Form features, such as bore and hole, can be used by the mechanical designer as modeling primitives and identified by feature extraction methods for reasoning about the manufacturability or other characteristics of a part. The idea of using features for processes other than machining is a recent research topic. As a result, a wide range of definitions and implementations have been presented for features. Due to the diversity of research in this area, the term feature, like many other terms in CAD, has multiple interpretations. In some CAD systems features are nothing but parameterized macros of solid primitives, containing only low-level geometric information. Other systems exploit the full potential of features for modeling not only geometric data but also topology, designer intent, engineering functions, and other information needed in an intelligent CAD system [3]. Three methods have been identified for creating the feature content of a product model [15]. In the first method, the feature content is extracted from the part geometry and/or topology [16-18]. In the second method, called design-with features [4,19,20], designers use features as modeling primitives. In the third method, designers interactively identify specific features in a model and attach attributes to geometric elements [21]. This paper concentrates on the methodology of design-with features and draws heavily on Dixon's extensive work in this area [4,19,20]. Dixon has proposed an architecture of a designwith-features system [4]. In this architecture, shown in Fig. 1, the designer can build a computer model of an artifact by using a library of design-with features and a set of add, modify, and delete operations. During the modeling process a monitor ensures that the designer's requested operations for manipulating selected features are allowed and understood by the system. Subsequently, the primary representation of the design, which is composed of design-with features, is automatically converted to lower-level abstractions that are referred to as secondary representations. These secondary representations, such as geometry or topology, are created for the modules of the system and used subsequently by those modules to reason about the specific design characteristics. Dixon emphasizes the system's ability to construct the secondary representations from the user-created primary represen-
A Feature-Based Approach to Structural Design
3
9 primary
representatlon~~
" interface library monitor
l.0 2N
~ c]l I~N~ I~raphi ,
,
t
redesign suggestions other
convertors Fig. 1. Architecture of a design-with-features system (after [4]).
tations as the major key to a successful implementation of a design-with-features system. Moreover, the dependencies between the primary and the secondary representations further constrain the nature of design-with features provided in the library. The design-with features are categorized into different classes, and each class is decomposed into types. For example, primitive and macro are types of features in the class of static features. Each type also contains specific modeling characteristics, for example, intersection features specify the details of how primitive and add-on features are connected. The attributes of features are a function of their type, so the primary representation of an artifact reveals the kind of information, such as geometry and topology, of the design. Depending on the sophistication of features used, this information can also contain designer intent and/or part functions. This type of information can potentially facilitate the interpretation and reasoning processes at later stages of design and redesign. Although a design-with-features system has many advantages over the existing CAD systems, such as facilitating the modeling process with the use of high-level application-specific primitives and providing better data representation methods to capture types and levels of design information, it also imposes certain limitations on designers. The most severe of these limitations is the finite number of
features and operations provided in the libraries of the system. The designer may need special purpose modeling primitives for representing some artifacts that otherwise cannot be modeled with the standard library features. One solution to this problem is to allow designers to define their own user-defined features and to supply the necessary supporting functions needed by the system such as converting primary representation of user-defined features into secondary ones. Another solution is to offer a set of generic, parameterized feature-primitives that can be assembled according to the designers' specific needs. The latter solution requires no implementation effort from the users, whereas the former provides more versatile and powerful tools to sophisticated users with programming expertise. Design-with features are primarily suited for use as modeling primitives in a top-down approach, although low-level features can be assembled bottomup to create high-level features. This top-down characteristic is strongly supported by the ability of features to represent high-level abstractions, such as functionality and designer intent, that are required during the preliminary design processes. However a different class of features that use a bottom-up approach to design has been proposed by Woodbury [22]. In this approach design information is associated with primitive components of a solid model, such as faces and edges, referred to as features. The designer who uses these features first builds a solid model of the part and then associates the design information with the particular components of that model. This approach is similar to the traditional methods offered by most CAD systems, except that the nongeometric design information is directly associated with the solid model of the part. Thus the system can use this data directly from the solid representation of the model to reason about the geometry and other design information of an artifact. Although the bottom-up features provide useful properties, such as aggregation of hierarchy and uniform interface to geometric objects, they do not have the rich and powerful vocabulary of higherlevel descriptors offered by the top-down approach. Users of the bottom-up features still must have detailed knowledge about the low-level components of the design in order to create the higher-level elements; therefore, these features are not suitable for the preliminary design process. The approach presented in this paper is based on design-with features and is intended to facilitate the top-down design process. However, the idea of user-defined features and macro facilities that are used during the detailed
4
M.K. Zamanian et al.
design process resembles the bottom-up design methodology.
3 Feature-Based Structural Design In this section we discuss the use of features for structural design. First, the process of structural design and the primary components used in this process are reviewed. Then the use of specific design-with features as the modeling components of this process are examined. The main focus of this study is on the structural design of buildings because this type of design encompasses a large and diverse class of structures; however, the ideas presented here are sufficiently general to be extended for other structural design problems.
3.1
Structural Design
The process of structural design begins with the specification of the functional requirements of a structure such as a building or a bridge. These requirements are then broken down into subfunctions that can subsequently be satisfied by structural components of different forms. These structural components may in turn be decomposed into more detailed subcomponents that perform more specific structural functions. For example, the preliminary functional specification of an office building may state that the building must accommodate up to 1500 people, have parking space for 500 automobiles, be no taller than 10 stories high, and occupy no more than an area of 180 by 100 meters in a specific downtown location. Architects prepare preliminary plans for a building to satisfy the above functional specifications as well aesthetic requirements. Generally, these plans consist of the spatial arrangements of the components of the building, such as rooms, stairwells, and corridors, along with geometric attributes of some of the high-level structural components, such as the frame and floors. After approval by the owner of the building the architectural specification of the building is submitted to the structural firm. The structural design firm then decomposes the building into several high-level components, such as frame, foundation, and floor systems, and a specialized design team is assigned to each of these components. When necessary, each team further decomposes the high-level components into lowlevel structural components, such as beams, columns, and baseplates. These low-level components are then designed and analyzed to satisfy the
Structural
Architect
Designer Building
Foundation
Frame
I t baseplates t girders footings columns piles braces
t mainbeams ~
I walls
cross beams r- stairs slabs L roof
Fig. 2. The design process and abstractions for a structural building.
spatial, structural, and code requirements. Figure 2 presents a summary of the design process described here. The process described previously is applicable to almost all types of structural design. One can observe that this process follows a top-down design methodology, that is, as the design progresses, the granularity of components decreases. Thus CAD tools ought to follow the same top-down methodology. However, this is not the case for a majority of existing CAD systems provided to structural designers. These systems are primarily used for drafting the nearly finished structures, and therefore do not play a role during the preliminary design or the redesign processes. To take advantage of the full potential of CAD technology throughout the design process, structural designers should be able to use intelligent CAD systems developed for specific structural applications. This will enable structural designers to deal with the components of a structure at different levels of abstraction that are suitable for a particular stage in the design process. Moreover, these designers will be able to use more powerful modeling primitives that can capture not only the three-dimensional geometric data but also specific information about the topology, functionality, and designerintent of the parts of a structure. The ability of feature-based design systems has been demon-
A Feature-Based Approach to Structural Design
strated in many applications of mechanical design, and we believe that structural CAD systems will benefit from this approach as well.
3.2 Design-with Features for Structural Design Two primary benefits are associated with designwith features: They provide users with high-level, specialized, modeling primitives that facilitate a top-down design process; and they contain a variety of design information that enable application programs to reason about the characteristics of a structure. Other researchers have explored the use of engineering data objects, similar to the data abstractions of object-oriented programming languages, for the representation of data and knowledge pertaining to structural designs [6,23,24] and electrical designs [25]. The feature-based approach presented in this paper also uses the concept of data objects to provide direct mapping to components in an engineering system [23]. Design-with features are specialized modeling entities that satisfy the specific needs of various design disciplines; however, the primary attribute of each design-with feature is still its three-dimensional geometry. In this section these ideas are explored for structural design in general and for the more specific case of building design. Structures are not monolithic but are made up of many substructures, components, and members. In this respect they are similar to mechanical assemblies with the advantage that each component is selected to perform a particular function. In addition, structural assemblies follow a design formalism. For a majority of structures this design formalism is routine; that is, the designer follows established guidelines to assemble standard components to produce a structure satisfying specific structural functions. Many complex structures require some nonstandard components as well as some nonroutine design and fabrication techniques, but because routinely designed and fabricated structures require less time and money, the number of nonstandard components is kept to a minimum. In fact, many of the bridges in the United States are designed and constructed based on existing templates that have been developed by major structural firms in cooperation with the Federal Highways Administration. The standard components and the formalism of the design and fabrication processes used in routine structural design are the primary motivations for using design-with features. While nonroutine design issues are not addressed in this paper, the use of
5
user-defined and macro features, discussed in section 3.2.4, are also relevant to issues in nonroutine design.
3.2.1
Hierarchy of structural features
As part of a larger effort to establish an international standard for the exchange of product information [26], several research groups have been working on the development of data models for the formal representation of structures such as buildings [24,27,28] and ships [29]. These models use an abstraction hierarchy approach [1] to decompose a structure into groups of functionally related data. The components of each group can be further decomposed into low-level entities until the desired granularity and characteristics of the design have been specified. As discussed earlier, the top-down design process of structures follows this same decomposition methodology. Figure 2 presents the top-down design process and the abstraction hierarchy of a building structure. To take advantage of the hierarchical nature of the structural design process and the abstractions used in this process, we propose a hierarchy for designwith features for structures. At the early stages of the design, often referred to as the conceptual design process [7], designers are mainly concerned with the global representation of a structure and not with the specific details of individual components. At this stage designers must be able to use highlevel features that effectively hide details about the tow-level components while emphasizing the attributes important to the structure's global scheme. For example, a high-level feature representing a frame should contain attributes for its type, number of stories, number of bays, orientation, and overall spacings and dimensions. This information is sufficient for reasoning about the behavior of the frame in relation to other high-level features of the structure. Once the conceptual design of the structure is completed, the detailed design of the individual members and their connections begins. Next, the detail designers assemble instances of low-level structural features that satisfy the functional requirements of the parent components. Moving down this hierarchy during the design process, the specified attributes of the parent features impose constraints on the features that descend from them. Satisfying these constraints ensures the integrity of the structure. For low-level features that are shared by several high-level features, such as a column that is part of both the gravity and the lateral systems, the constraints imposed by all the parent features must be satisfied. Different secondary representations of
6
these features may be necessary to provide suitable mechanisms for evaluating the constraints imposed by different parent features. One of the most important and complex elements of the component hierarchy of a structure is the connection between components. A connection differs from other types of structural components in that it is not a stand-alone entity, and its attributes are dependent on the members that it connects. Although some attributes of a connection can be categorized into parametric entities, such as bolts, holes, and welds, customized parts of a connection, such as cut-out sections, cannot be dealt with in this manner. Therefore, a connection is represented as a special case of an abstract feature whose attributes correspond to its own elements as well as to the relevant characteristics of the members it connects.
3.2.2 Attributes of structural features The attributes of each feature depend on the specific application. For structural design we have chosen the following attributes to be common to all features: 9 Geometry 9 Topology Structural functionality 9 Standard provisions 9
The information required to specify these attributes is either provided with the feature and/or requested from the designer. For example, a feature may have its structural functionality specified by the system as "support gravity load," but the designer may have to provide values of the parameters needed for its geometrical representation. In either case, constraints may be associated with the values of these attributes. Conformance to these constraints can be enforced by the monitor and/or application modules of the design-with-features system. Depending on the level of detail necessary to define the geometry of a feature, the geometry attribute may range from a simple stick-figure to a complex solid model. High-level features used during the conceptual design process usually do not require detailed geometric specifications; thus simple stick-figure diagrams are sufficient to reveal the overall sizes and topology of these features. However, detailed representations of low-level features, which correspond to the structural members and connections, are necessary for design and fabrication of these members. This detailed, three-dimensional representation can be in the form of a parametric solid model. For extruded shapes, such as wide-flange beams or slabs, concise and accurate
M.K. Zamanian et al.
representation can consist of a parameterized cross section and the length of the member. For irregularly shaped members, more complex solid representation generated by CSG or B-rep may be necessary. In addition, constraints may be imposed on the parameters defining the geometry and component relationships in order to ensure that design requirements are satisfied. For example, width/thickness limitations are established by design specifications to ensure that overall column buckling rather than local buckling governs the member strength [30]. The connectivity and orientation information of a structural feature is stored in its topology attribute. A structural feature is usually connected to one or more other features, and often only specific types of features can be connected together. For example, a slab can be connected to several beams in a floor system, but a baseplate of the foundation may never be connected to a slab. Furthermore, some structural features can be connected to others only at specific locations and/or with special connection methods to satisfy their structural functionality. Similarly, the spatial orientations of some structural features affect their functionality and behavior; for example, the section moduli of wide-flange beams depend on the orientation of their loading plane. The functionality of a structural feature is either specifically predetermined in the system or entered as part of the designer-intent information. For instance, the structural functionality of a baseplate is to distribute a load to masonry within the allowable bearing capacity of the concrete. On the other hand, a beam may be used for a variety of structural functions, such as transferring of gravity loads in a floor system or supporting lateral loads as part of a rigid frame. In the case of more complex structures a feature may provide seve~_J different structural functions, some of which are the result of its special arrangement with other features. In such cases the functionality attribute may consist of an ordered set of structural functions. For example, a floor system is designed primarily to support the gravity loads; however, it also serves as a diaphragm that resists transverse shear forces. Thus the first and second entries in the functionality attribute set of this feature may be specified as "support gravity load" and "resist transverse shear," respectively. Almost every component of any structure must be designed to meet specific standard provisions. Most of the existing provisions are for the detailed design and analysis of the low-level structural members [30,31], but some provisions apply to the high-level components of structures, for example, building fire and occupancy codes, structural requirements for
A Feature-Based Approach to Structural Design
seismic loading, and bridge load limits. Recently, a number of knowledge-based programs have incorporated the applicable standard provisions for specific domains of structural design in their systems [6,32]. In the proposed feature-based system this type of information could be included in each feature's standard provisions attribute. This approach provides the necessary information needed by the monitor and the application programs for checking the compliance of a feature with the specifications; however, it is not meant to serve the needs of a general standards processing program. The attributes described earlier are common to all design-with features used for structural design. More specific attributes are needed to describe the characteristics of those components used in particular structural applications. For example, the feature used for a concrete column must contain the specific information about the arrangement and type of reinforcement bars as well as more general, geometry, and topology attributes. Presenting the specific attributes of structural features requires the detailed study of the characteristics of these features and is beyond the scope of this paper.
3.2.3 Operations for structural features Three main operations have been suggested by Dixon [4] to manipulate design-with features: create, modify, and delete. These operations are sufficiently general for the proposed structural features. The create operation allows designers to make an instance of a design-with feature by specifying its required attributes. Once a feature has been created, its attributes can be altered with the modify operation, or it can be removed from the design by the delete operation. We suggest two additional operations for structural features: expand and query, respectively. The former operation is used to expand the high-level conceptual features into an assembly of low-level structural features. This operation is particularly useful when a detail designer needs a concrete realization, an arrangement of specific low-level features, for an abstract feature, that is, for example, a floor system, that was created by a conceptual designer. The macro and template facilities, discussed in section 3.2.4, also benefit from this operation. The query operation enables designers to obtain information about the available features as well as the attributes of feature instances used in the design. 3.2.4 User-defined features and macro facilities As mentioned earlier, one of the drawbacks of a simple feature-based CAD system is the limited number of features in its library. In the domain of
7
structural design some design requirements cannot be satisfied with available standard components. Consequently; built-up sections must be designed and fabricated. However, built-up sections are rarely used only once in a large structure. Designers often have a selection of parameterized, built-up sections available for similar designs. The ability to add user-defined features not only offers better modeling capabilities to designers but also provides a useful technique for abstracting design information. The process of creating user-defined features can be facilitated with the use of the macro facilities. Macro facilities provide formal methods for combining the features in a library. The simplest form of macro facilities is a feature template that is arranged according to the specific needs of a design problem. For structural members that cannot be composed from available features in the library, designers may have to implement their own features. This implementation can potentially require an extensive knowledge of the system, and therefore is only recommended for sophisticated users. Some of the general high-level features, such as frames and floor systems, will be provided in an abstract form in the library of the system and can be used directly during conceptual design. Designers may also need to create their own high-level features or to create an assembly of low-level structural members that can be expanded to transform a high-level feature into an assembly. The user-defined and macro facilities can assist designers in these tasks more easily and efficiently. These facilities will allow designers to construct any desirable structural component in a bottom-up fashion while the overall schema of the design benefits from a topdown methodology.
4
Conclusions and Future Directions
We have presented a new approach using features for structural design. The primary motivation for this work is the need for intelligent structural CAD systems that make better use of information modeling techniques to facilitate a top-down design process. Current CAD systems for structural design lack the diverse modeling and reasoning capabilities needed throughout the design process; they are drafting systems with advanced computer graphics and data management facilities whose primary goal is to facilitate the modeling of nearly finished, detailed designs of structures. In order to develop CAD systems that can be used for conceptual as
8
M.K. Zamanian et al.
well as detailed design, we need primitives more sophisticated than lines and annotations. The current research on the use of design-with features in several applications of mechanical design has produced promising results for development of a new generation of intelligent CAD systems. We believe that similar principles can be applied in the domain of structural design because most structures can be viewed as assemblies of components that possess specific shapes and functions. The primary attribute of each structural feature is its three-dimensional geometry, but other design attributes, such as topology and functionality, are stored with the geometry. Depending on a feature's type and usage, its attributes may be predefined and/or specified by the designer. The values of these attributes are monitored by a special module in the CAD system to assure that all design constraints and preconditions are satisfied. The primary representation of the structure, which is composed of features, is converted to several low-level representations that in turn are used by application programs for activities like display, cost estimation, and so on. The work presented in this paper is part of a larger effort at Carnegie Mellon University for research and development of better computer-aided tools for design and formal representation of structures. Current work focuses on designing the necessary abstractions for formal representation of structural components of constructed facilities such as buildings. The results of this work will be used to define the attributes of a design-with-features system for constructed facilities. References 1. Eastman, C.M. (1982) Recent developments in representation in the science of design. Technical Report DRC-15-1182, Design Research Center-Carnegie Mellon University, Pittsburgh 2. Gossard, D.C.; Zuffante, R.P.; Sakurai, H. (1988) Representing dimensions, tolerances, and features in MCAE systems. IEEE, 2, 51-59 3. Rossignac, J.R.; Borrel, P.; Nackman, L.R. (1988) Interactive design with sequences of parameterized transformations. In: Intelligent CAD Systems 2: Implementational Issues. (Eds. P. ten Hagen, T. Tomiyama) New York: Eurographics, Springer-Verlag 4. Cunningham, J.J.; Dixon, J.R. (1988) Designing with features: The origin of features. In; Computers in Engineering (Eds. V.A. Tipnis, E.M. Patton). San Francisco: ASME, pp. 237-242 5. Shah, J.J.; Rogers, M.T. (1988) Expert form feature modeling shell. Comput. Aided Des. 20(9), 515-524 6. Maher, M.L. (1985) HI-RISE: A knowledge-based expert system for the preliminary structural design of high rise
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
buildings. PhD thesis, Carnegie Mellon University, Pittsburgh Fenves, S.J.; Baker, N.C. (1987) Spatial and functional representational language for structural design. In: Expert Systems in Computer-Aided Design. IFIP 5.2. New York: Elsevier Science (North-Holland) Woodbury, R.F.; Oppenheim, I.J. (1987) An approach to geometric reasoning. In: WG 5.2 First International Workshop on Intelligent CAD, Cambridge, MA. October 6-8 Jared, G.E. (1984) Shape features in geometric modeling. Solid Model. Comput. 121-137 Birchfield, B.; Downey, P. (1984)Product definition data interface. In: Advances in Distributed Data Base Management for CAD/CAM, Denver. NASA Conference Publication 2301 Faux, I.D. (1986) Reconciliation of design and manufacturing requirements for product description data using functional primitive part features. Technical Report R-86-ANC/ GM/PP-01, CAM-I, Arlington, TX Henderson, M.R.; Anderson, D.C. (1984) Computer recognition and extraction of form features: A CAD/CAM link. Comput. Indust. 6(4); 315-325 Miner, R.H. (1985) A method for the representation and manipulation of geometric features in a solid model. Master's thesis, MIT, Cambridge, MA Pratt, M.J. (1984) Solid modeling and the interface between design and manufacture. IEEE Comput. Graph. Appl. July, 52-59 Chung, J.C.; Cook, R.L.; Patel, D.; Simmons, M.K. (1988) Feature-based geometry construction for geometric reasoning. In: Computers in Engineering (Eds. V.A. Tipnis, E.M. Patton) San Francisco: ASME, pp. 497-504 Henderson, M.R. (1984) Extraction of feature information from three dimensional CAD data. PhD thesis, Purdue University, West Lafayette, IN Woo, T.C. (1982) Feature extraction by volume decomposition. In: Conference on CAD/CAM Technology in Mechanical Engineering. Cambridge, MA, pp. 76-94 Pinilla, J.M.; Finger, S; Prinz, F.B. (1989) Shape feature description and recognition using an augmented topology graph grammar. In: Engineering Design Research Conference, Amherst, MA. NSF Libardi, E.C.; Dixon, J.R.; Simmons, M.K. (1986) Designing with features: Design and analysis of extrusion as an example. In: Design Engineering Conference, Chicago, March 24-27, ASME Luby, S.C.; Dixon, J.R.; Simmons, M.K. (1986) Designing with features: Creating and using a features database for evaluation of the manufacturability of castings. In: Computers in Engineering Conference, Chicago, ASME CAM-I Inc. (1986) Dimensions and tolerances feasibility demonstration. Technical report, CAM-I Inc., Arlington, TX Woodbury, R.F. (1987) The knowledge-based representation and manipulation of geometry, PhD thesis, Carnegie Mellon University, Pittsburgh Fenves, G.L. (1989) Object oriented models for engineering data. In: Computing in Civil Engineering, Atlanta, GA: ASCE, pp. 564-571 Powell, G.H.; Bhateja, R. (1988) Data base design for computer-integrated structural engineering. Eng. Comput. 4, 135-143 Katz, R.H. (1985) Information Management for Engineering Design. New York: Springer-Verlag
A Feature-Based Approach to Structural Design
26. Warthen, B. (1988) P D E S - - A CAD standard for data exchange. Unix World December, 103-104 27. Bjork, B. ; Penttila, H. (1989) A scenario for the development and implementation of a building product model standard. In: Current Research and Development in Integrated Design, Construction and Facility Management. Stanford, CIFE, Stanford University 28. Lavakare, A.; Howard, H.C. (1989) Structural steel framing data model. Technical report, Center for Integrated Facility Engineering-Stanford University, Stanford 29. Corn, E.J.; Gerardi, M.L.; Vander Schaaf, J.R. (1988) Ref-
9
erence model for ship structural systems. Technical Report ISO TC184/SC4/WG1 N188 Version 3.0, The National Shipbuilding Research Program 30. American Institute of Steel Construction. (1980) Steel Construction, 8th ed. American Institute of Steel Construction, Inc., Chicago 31. American Concrete Institute. (1987) Manual of Concrete Practice. American Concrete Institute, Detroit 32. Garrett Jr., J.H.; Fenves, S.J. (1987) A knowledge-based standards processor for structural component design. Eng. Comput. 2, 219-238