Joshua Shaffer

Benchmarking CAD search techniques

While benchmark datasets have been proposed for testing computer vision and 3D shape retrieval algorithms, no such datasets have yet been put forward to assess the relevance of these techniques for engineering problems. This paper presents several distinctive benchmark datasets for evaluating techniques for automated classification and retrieval of CAD objects. These datasets include (1) a dataset of CAD primitives (such as those common in constructive solid geometry modeling); (2) two datasets consisting of classes generated by minor topological variation; (3) two datasets of industrial CAD models classified based on object function and manufacturing process, respectively; (4) and a dataset of LEGO© models from the Mindstorms© robotics kits. Each model in the datasets is available in three formats - ACIS SAT, ISO STEP, and as a VRML mesh (some models are available under several different fidelity settings). These are all available through the National Design Repository.Using these datasets, we present comprehensive empirical results for nińe (9) different shape and solid model matching and retrieval techniques. These experiments show, as expected, that the quality of precision-recall performance can significantly vary on different datasets. These experiments reveal that for certain object classes and classifications, such as those based on manufacturing processes, all existing techniques perform poorly. This study reveals the strengths and weaknesses of existing research in these areas, introduces open challenge problems, and provides meaningful datasets and metrics against which the success of current and future work can be measured.

Workflow for integrating mesoscale heterogeneities in materials structure with process simulation of titanium alloys

In this paper, a generalized workflow is outlined for the necessary integration of multimodal measurements and multiphysics models at multiple hierarchical length scales demanded by an Integrated Computational Materials Engineering (ICME) approach to accelerated materials development. Recognizing that multiple choices or techniques are typically available in each of the main steps, several exemplary analyses are detailed utilizing mainly the alpha/beta titanium alloys as an illustrative case. It is anticipated that the use and further refinement of these workflows will promote transparency and engender intimate collaborations between materials experts and manufacturing/design specialists by providing an understanding of the various mesoscale heterogeneities that develop naturally in the workpiece as a direct consequence of the inherent heterogeneity imposed by the manufacturing history (i.e., different thermomechanical histories at different locations in the sample). More specifically, this article focuses on three main areas: (i) data science protocols for efficient analysis of large microstructure datasets (e.g., cluster analysis), (ii) protocols for extracting reduced descriptions of salient microstructure features for insertion into simulations (e.g., regions of homogeneity), and (iii) protocols for direct and efficient linking of materials models/databases into process/performance simulation codes (e.g., crystal plasticity finite element method).

CAD Archives Based on OAIS

Within the past few years, there has been a steady, substantial growth of interest in “long-term” archiving of digital data. This problem is particularly acute in many branches of engineering design, where cycles of technological obsolescence in supporting tools happen much more rapidly than those of designed products. Capturing and preserving design knowledge through these cycles is a major challenge that has come to be recognized by many government, industry, and research organizations. The ability to do so has important operational, efficiency, and legal ramifications for the manufacturing industry and its customers. This paper describes this problem, presenting examples of both why it must be addressed and why it is a challenge. In particular it relates preservation of engineering data to digital archiving efforts in other domains as well as ongoing work within the engineering research community on design repositories. As is shown, long term archiving of digital design knowledge draws upon both but possesses its own unique issues. Much of this discussion is couched within the language of the ISO Open Archival Information Systems (OAIS) Reference Model, including a mapping from an existing significant design repository into the OAIS model. In this way, it is hoped that this paper will widen the discussion on digital archiving within the community of this conference as well as help connect to research in other areas.© 2006 ASME

Web Service Interfaces for Design Repositories

Reuse of design knowledge is an important goal in engineering design, and has received much attention. A substantial set of algorithms, methodology, and developed systems exist which support various aspects of this goal. However, the majority of these systems are built around a particular user interface, often some form of Web-based repository portal. The work described here presents search and other core functionality as web services rather than a monolithic repository system. These services may then be employed by a variety of applications, integrating them into interfaces familiar to the designer, extending functionality, streamlining their use, and enabling them to be employed throughout the design process. This paper demonstrates this approach by wrapping previously developed repository search algorithms as web services, and then using these within a plug-in for an existing commercial CAD environment. Based on issues encountered in developing this demonstration, this paper also discusses the challenges and potential approaches toward a more general, widespread application of web services in engineering design.Copyright

Microstructure-Informed Cloud Computing for Interoperability of Materials Databases and Computational Models: Microtextured Regions in Ti Alloys

With the fast global adoption of the Materials Genome Initiative (MGI), scientists and engineers are faced with the need to conduct sophisticated data analytics on large datasets to extract knowledge that can be used in modeling the behavior of materials. This raises a new problem for materials scientists: how to create and foster interoperability and share developed software tools and generated datasets. A microstructure-informed cloud-based platform (MiCloud™) has been developed that addresses this need, enabling users to easily access and insert microstructure informatics into computational tools that predict performance of engineering products by accounting for microstructural dependencies on manufacturing provenance. The platform extracts information from microstructure data by employing algorithms including signal processing, machine learning, pattern recognition, computer vision, predictive analytics, uncertainty quantification, and data visualization. The interoperability capabilities of MiCloud and its various web-based applications are demonstrated in this case study by analyzing Ti6AlV4 microstructure data via automatic identification of various features of interest and quantifying its characteristics that are used in extracting correlations and causations for the associated mechanical behavior (e.g., yield strength, cold-dwell debit, etc.). The data were recorded by two methods: (1) backscattered electron (BSE) imaging for extracting spatial and morphological information about alpha and beta phases and (2) electron backscatter diffraction (EBSD) for extracting spatial, crystallographic, and morphological information about microtextured regions (MTRs) of the alpha phase. Extracting reliable knowledge from generated information requires data analytics of a large amount of multiscale microstructure data which necessitates the development of efficient algorithms (and the associated software tools) for data recording, analysis, and visualization. The interoperability of these tools and superior effectiveness of the cloud computing approach are validated by featuring several examples of its use in alpha/beta titanium alloys and Ni-based superalloys, reflecting the anticipated computational cost and time savings via the use of web-based applications in implementations of microstructure-informed integrated computational materials engineering (ICME).

Intellectual teamwork on smartphones

We demonstrate a software system that runs on smartphones and allows two or more geographically dispersed participants to collaborate on the solution of mathematical problems. We show how participants can create and join a collaborative session, how they can use a virtual whiteboard for the exchange of geometrical drawings and of mathematical formulas, and how they can communicate using text messages. We demonstrate the turn-management mechanism provided by our system, and we show how the system facilitates cross-referencing within the session.

Mobile mathematics communication

We present a system [1] that allows wireless smartphones to beused for mathematics communication, that is, for the creation andexchange of mathematical formulas, diagrams, and text between twoor more participants. The system enables two or more persons withsmartphones or traditional computers to participate in a session.Each of the participants may convey textual, graphical andmathematical information to the other participants. Users can draw,edit, and label geometric shapes, send chat messages, and composeformulas. A turn taking mechanism moderates the communication. Thesystem also supports the integration of services that can be usedto provide individual users with additional functionality.Currently, a LATEX rendering service is available to allow users tocreate and share mathematical formulas in typeset quality. Wireless smartphones are becoming the medium of choice forimprovised synchronous collaboration since increasing numbers ofusers carry their smartphones at all times. It is true that thesmall size of the devices---while necessary for theirubiquity---limits the complexity of collaborative tasks that can becarried out effectively. On the other hand, there is a need tocapture inspiration, to access and evaluate information on the go,and to make decisions on the spot. The domain of mathematics is ideally suited to explore---andpush---the limits of smartphone communication. The challenge ofrepresenting mathematics in typeset form has led to the developmentof document preparation systems such as TEX, LATEX, and LYX whichare in widespread use today. Mathematical handwriting recognitioncontinues to push the limits of general handwriting recognition[2]. Many of the cognitive challenges that arise in mathematicalcollaboration also arise in intellectual teamwork in otherdomains. Some of the challenges of developing a system for mobilemathematics communication are posed by the input and outputlimitations of the devices and the enormous heterogeneity of theavailable hardware platforms. We use the Treo 600/650 as our hardware platform. Our softwareis written in pure Java and can be run on any platform thatsupports the Java Virtual Machine. Our software architecture makesit easy to add new local or remote services to enhance thecommunication. The poster presents use-case diagrams, architecture diagrams,and a series of screenshots that describe the users interactionwith the system. The poster also shows how we solved the problemsof cross-referencing and turn-taking in mobile collaboration. A7.5-minute video shows users interacting with the system.