Steven R. LeClair

THE RAMAN SPECTRUM OF TI3SIC2

In this paper, the Raman spectrum of Ti3SiC2 is reported and compared with that of TiC0.67. All the TiC0.67 first order Raman disorder-induced modes are active, but shifted, in Ti3SiC2. Two additional peaks at 150 and 372 cm−1 are observed in Ti3SiC2. The former is ascribed to a shear mode between the Si and Ti planes; the origin of the latter is unknown. No second order Raman bands are detected. Micro-Raman spectroscopy also reveal the presence of ≈50 A graphite crystallites in samples hot pressed in graphite dies—these crystallites are not detected in samples processed by hot isostatic pressing in molten glass containers.

Integration of design and manufacturing: solving setup generation and feature sequencing using an unsupervised-learning approach

Abstract An approach to the integration of design and manufacturing knowledge is presented that uses a feature-based design environment and an unsupervised learning algorithm to categorize features into a setup for machining. Setup generation is a pivotal step in automated process planning as it greatly influences machine/tool selection, machining sequences and fixture configuration. The proposed algorithm and architecture incorporate multiple objective functions into setup generation. Intersecting and nonintersecting features within a setup are identified and classified using an associative memory. After the feature sequence has been determined, several algorithms are proposed to obtain the best tool sequence for creating the features in a setup.

Rough sets applied to the discovery of materials knowledge

The functional mapping of material structure to properties, processing, and use is the principal driver for all scientific and engineering endeavors. Given the high cost of experimentation and the computational intractability of ab initio materials research, more efficient and accurate predictions of yet-to-be-made materials is an equally prominent endeavor, if not a preeminent materials research frontier. Because of the vast amounts of information to be considered in the pursuit of either, the automation of search-based methods for augmenting more analytic approaches is receiving increasing attention. Given the computational challenges to automation and to retrieving data from complex databases, search-based methods offer an expeditious approach to providing a researcher both insight and perspective. Rough sets is discussed relative to these objectives, as is current research to address its limitations and difficulties in application. Several materials related examples are offered to illustrate the application of the method.

An incremental adaptive implementation of functional-link processing for function approximation, time-series prediction, and system identification ☆

Abstract This paper presents an adaptive implementation of the functional-link neural network (FLNN) architecture together with a supervised learning algorithm that rapidly determines the weights of the network. The proposed algorithm is able to achieve ‘one-shot’ training as opposed to iterative training algorithms in the literature. Also discussed is a stepwise updating algorithm that updates the weights of the network while importing new observations. The proposed algorithms have also been tested on several data sets and the simulation shows a very promising result.

Qualitative Process Automation

Abstract This paper addresses research in computer science, control theory and material processing science wherein the focus is to develop a process control system which enables in situ process model development. A generic control system architecture employing the use of qualitative physics (QP) is described together with an explanation of the system operation involving multiple co-operating expert systems. Specific details involving development and application of the control system to auotoclave curing of composites is also presented, together with recent research and production results.

Feature sequencing in the rapid design system using a genetic algorithm

This paper addresses the feature sequencing problem in the Rapid Design System (RDS). The RDS is a feature-based design system that integrates product design and process planning. An important issue in feature-based process planning for machined parts is the order in which material is removed to form the resultant part. The order, or sequence, is partially dependent on the geometric relationships between features. The sequence affects the safety, the time it takes to machine the part, and the quality of the finished part. The sequence of material removal depends on two types of relations between features: (1) intersections and (2) interfeature associations. Both types of relations compound the search for an ‘optimal’ sequence of material removal. Therefore, the research problem has been the discovery and development of a genetic algorithm (GA) that efficiently searches the solution space for all possible sequences and identifies the best sequences in terms of safety, time and quality.

Materials structure-property prediction using a self-architecting neural network

An important trend in materials research is to predict properties for a new material before committing experimental resources. Often the prediction is motivated by the search for a material with a unique combination of properties. The selection of a property or feature is crucial to the plausibility of the prediction. This paper proposes the use of a self-architecting neural network to model the relation between materials structure and properties for the purpose of predicting the properties of new materials, i.e. to predict properties for an unknown compound. In this paper, we summarize the prediction attained with the proposed neural network structure referred to as the Orthogonal Functional Basis Neural Network (OFBNN). The OFBNN, which combines a new basis selection process and a regularization technique, not only gives us a more computationally tractable method, but better generalization performance. Simulation studies presented here demonstrate the performance, behavior and advantages of the proposed network.

Local grain orientation and strain in polycrystalline YBa2Cu3O7−δ superconductor thin films measured by Raman spectroscopy

We report direct measurements of local grain orientation and residual strain in polycrystalline, C-axis oriented thin YBa2Cu3O7−δ superconducting films using polarized Raman spectroscopy. Strain dependence of the Ag Raman active mode at 335 cm−1 was calibrated and used to measure local strain in the films. Our data showed that high quality films are associated with the connected path of uniform grain orientation (single crystal-like) across the film and uniform residual strain in the range of −0.3%. Nonuniform grain orientation or high angle grain boundaries and nonuniform local strains were associated with low quality films.

Non-destructive, In situ Measurements of Diamond-Like-Carbon Film Hardness using Raman and Rayleigh Scattering

Diamond-like Carbon (DLC) coatings have recently proven to be suitable for a number of tribological applications. Hardness of the DCL coating is very important for such application. An in-situ, non-destructive technique to measure the film hardness would be crucial for process control and quality control of DLC coatings. In this study, Raman and Rayleigh scattering were investigated as potential techniques for non-destructive measurements of DLC film hardness. Features in both Raman and Rayleigh spectra were correlated with film hardness as measured by nano-indentation technique. Regarding the Raman spectra of the film, a linear correlation was found between film hardness and area under diamond related peak at 1332 cm−1. Regarding the Rayleigh scattering, a linear relationship was, also, found between the film hardness and the height of the Rayleigh line. Raman and Rayleigh scattering have shown huge potential to be used as non-destructive, in-situ techniques to measure DLC coating hardness. Copyright

Self-improving process control for molecular beam epitaxy

This paper addresses manufacturing research involving advances in material process control. The research objective has been to develop intelligent, self-directed and self-improving control systems which enablein situ (real-time) control path generation based on both product (material behaviour) and processing (control agent) feedback. A ‘product-process’ control philosophy which emphasises product quality is described together with a generic architecture for representing product and process knowledge.Specific details are presented involving the development and application of a self-directed and self-improving material processing system for molecular beam epitaxy of gallium arsenide wafers. Special emphasis is given to the development of a neural model for self-improving control as well as future research directions.