Robert G. Wilhelm

Task Specific Uncertainty in Coordinate Measurement

Abstract Task specific uncertainty is the measurement uncertainty associated with the measurement of a specific feature using a specific measurement plan. This paper surveys techniques developed to model and estimate task specific uncertainty for coordinate measuring systems, primarily coordinate measuring machines using contacting probes. Sources of uncertainty are also reviewed.

Adaptive sampling for coordinate metrology

An iterative sampling process for dimensional measurement is presented. The strategy is based upon the use of surface normal measurement data to develop an interpolating curve between sample points. The interpolant is used to select subsequent measurement targets iteratively. The process is repeated until the measurement converges to a complete and accurate evaluation of the surface. The required sample size is proportional to part quality. The most accurate parts will require the least sample points; whereas, lower quality parts will require a greater number of total samples. The method is particularly applicable to measurement of complex surfaces with coordinate measuring machines (CMMs).

Modeling and representation of geometric tolerances information in integrated measurement processes

Modeling and representation of geometric tolerances information across an enterprise is viable due to the advances in Internet technologies and increasing integration requirements from industry. In Integrated Measurement Processes (IMP), geometric tolerances data model must support different models from several well-defined standards: including ASME Y14.5M-1994, STEP, DMIS and others. In this paper, we propose a layered conformance level geometric tolerances representation model. This model uses the widely applied ASME Y14.5M-1994 as its foundation layer by abstracting most information from this standard. The additional geometric tolerances information defined by DMIS and STEP is incorporated into this model to form corresponding conformance layers that support IMP. Thus, different application domains in an enterprise can use this data model to exchange product information. This model is further transformed with XML Schema that can be used to generate XML instance file to satisfy geometric tolerances representation requirements in IMP.

Automating tolerance synthesis: A framework and tools

Abstract This paper describes CASCADE-T—a new approach to tolerance synthesis that uses a complete representation of the conditional tolerance relations that exist between features of a part under design. Conditional tolerances are automatically determined from functional requirements and shape information. Tolerance primitives based on the virtual boundary requirements approach to tolerance representation are composed to form more complex tolerance relationships. Artificial intelligence techniques, including a constraint network, frame-based system, and dependency tracking are used to support flexible and detailed computation for tolerance analysis and synthesis.

Part Form Errors Predicted from Machine Tool Performance Measurements

Abstract Machine tool performance testing, as defined by ISO 230 and ANSI B5.54, has been successfully used to maintain and improve the accuracy and repeatability of production-level machine tools. In this study, a controlled series of experiments have been used to test the efficacy of these performance tests in the prediction of part form errors. Results are shown for flatness, squareness, position, and profile tolerances. The experimental results suggest that standard machine tool performance tests can also be used to predict the “best-case” tolerances that can be achieved for particular part features.

Computer methods for tolerance design

This book describes recent research advances and computer tools that can be applied in the determination of geometric tolerances. A framework for tolerance synthesis is developed and used with artificial intelligence techniques to provide computer methods for both analysis and synthesis of geometric tolerance specifications. Tolerance primitives, based on a sound theory of tolerancing, are used to represent tolerance relationships or links between geometric entities and functional requirements. Algorithms are developed for the determination of boundedness and the measurement of sufficiency. A detailed constraint network is used to represent tolerance relations for a part under design and provide for the composition of tolerance specifications.

Integrating manufacturing softwares for intelligent planning execution: a CIIMPLEX perspective

Recent developments have made it possible to interoperate complex business applications at much lower costs. Application interoperation, along with business process re- engineering can result in significant savings by eliminating work created by disconnected business processes due to isolated business applications. However, we believe much greater productivity benefits can be achieved by facilitating timely decision-making, utilizing information from multiple enterprise perspectives. The CIIMPLEX enterprise integration architecture is designed to enable such productivity gains by helping people to carry out integrated enterprise scenarios. An enterprise scenario is triggered typically by some external event. The goal of an enterprise scenario is to make the right decisions considering the full context of the problem. Enterprise scenarios are difficult for people to carry out because of the interdependencies among various actions. One can easily be overwhelmed by the large amount of information. We propose the use of software agents to help gathering relevant information and present them in the appropriate context of an enterprise scenario. The CIIMPLEX enterprise integration architecture is based on the FAIME methodology for application interoperation and plug-and-play. It also explores the use of software agents in application plug-and- play.

Finite Schedule Monitoring and Filtering in a Computer Integrated Manufacturing Environment

Abstract An efficient and effective method of filtering is described for reactive scheduling in an integrated manufacturing environment. The filter evaluates the significance of schedule deviations in near-real-time and gives a statistical estimate of the importance of each schedule deviation; Highly important deviations being those that prevent schedule objectives from being achieved. A typical result from STC filtering would be an indication that there is a 95% probability of failing to meet the schedule if the resource were to continue working at the same rate. The technique is well suited for bottle-neck resources but may be applied to all resources in an enterprise. The filter, based on statistical throughput control (STC) is compared to past work in DEDS and reactive scheduling. Model formulation is presented to show extensions from previous applications of STC for short-term production control in a CONWIP (Constant WIP) production system. Data requirements are detailed with reference to Business Object Documents (BODs) defined by the Open Applications Group (OAG). A summary of the CIIMPLEX framework is presented to show how the filter is applied.

Genetic Algorithms for TTRS tolerance analysis

This paper describes a computational tool for TTRS tolerance analysis that makes use of a Genetic Algorithm approach. The Technologically and Topologically Related Surfaces (TTRS) concept represents and classifies part surfaces and mechanical assemblies. We show results for examples with position and orientation tolerances that suggest that a Genetic Algorithm approach can provide an effective means for tolerance analysis.

Electronic Commerce Negotiation in a Supply Chain Via Constraint Evaluation

Negotiation is of critical importance in e-commerce applications where the supply chain is dynamic and reconfiguring. In this research supply chain negotiation problems are addressed as constraint-satisfaction problems. In general each negotiation is handled in the largest scope possible to avoid the sub-optimality that can result from many local solutions. This global approach, however, must be balanced with time constraints that apply in e-commerce supply chain execution. In this paper, we describe a new approach for e-commerce supply chain negotiation via constraint evaluation. As well, results from prototype software, distributed across the internet, are discussed. Beyond the general formulation, we describe a more particular problem of kitted demand where a collection of purchased items must be acquired within the same time horizon. To address slow convergence a time-based penalty function is proposed.