Sencer Yeralan

Computerized roundness inspection

Computerized visual inspection systems are increasingly being considered and implemented as an alternative to manual inspection and manual gauging in manufacturing systems. Although computer vision systems and image processing have received much attention in the past, statistically powerful, yet computationally efficiency methods still remain to be developed to process the visual information to make computerized inspection systems cost effective. This study investigates the case of inspecting round production parts for shape and size. The statistical properties of the out-of-roundness measurement procedure are investigated. Real-time implementation considerations are discussed.

The minimax center estimation problem for automated roundness inspection

Abstract Significant advances in computer technology, sensing devices, pattern recognition and image processing techniques have made the on-line inspection of all production parts a feasible option in modern manufacturing systems. This study is concerned with the development of efficient algorithms to determine the out-of-roundness error for automated inspection of circular production parts. The out-of-roundness error is defined as the maximum deviation of a given sey of sample points, taken from the boundary of the part, from the so called minimax circle.

FINDING PLACEMENT SEQUENCES AND BIN LOCATIONS FOR CARTESIAN ROBOTS

Abstract We model the repetitive placement by a Cartesian robot of n parts on a rectangular workpiece. There are n bins or feeders (one per part), to be placed around the boundary of the workpiece, which contain the parts. The robot picks a part from a bin, places it, picks another part, places it, etc.; any placement sequence is possible. The problem, to find bin locations and a placement sequence to minimize total assembly time, is formulated as a traveling salesman problem (on a graph with n nodes) with special structure. This structure allows the computation of a lower bound on the minimum total assembly time in order n effort. The lower bound improves as n increases, and leads to a simple solution algorithm which gives asymptotically optimal solutions in order n log n effort. For die case where parts are uniformly distributed on the workpiece, we give simple closed-form expressions for the expected value of the lower bound. These expressions should be helpful for design decisions; for example, holdin...

Algorithm 980: Sparse QR Factorization on the GPU

Sparse matrix factorization involves a mix of regular and irregular computation, which is a particular challenge when trying to obtain high-performance on the highly parallel general-purpose computing cores available on graphics processing units (GPUs). We present a sparse multifrontal QR factorization method that meets this challenge and is significantly faster than a highly optimized method on a multicore CPU. Our method factorizes many frontal matrices in parallel and keeps all the data transmitted between frontal matrices on the GPU. A novel bucket scheduler algorithm extends the communication-avoiding QR factorization for dense matrices by exploiting more parallelism and by exploiting the staircase form present in the frontal matrices of a sparse multifrontal method.

Effect of buffer size on productivity of work stations that are subject to breakdown

A production line consisting of two work stations in series, and intermediate buffer of size m is modeled as a discrete-parameter Markov process with 4m + 8 states. Owing to the block structure of the transition probability matrix, the steady-state probabilities are obtained by the successive solution of systems of 4 simultaneous equations. Moreover, the probabilities of the interior states are shown to follow a scalar geometric relation Hence numerical solutions are very easily obtained even for large buffer sizes. An empirical formula for the production rate is presented. This formula is remarkably accurate over a wide range of parameters.

Genetic search with dynamic operating disciplines

Abstract This study extends the idea of genetic search to include the selection of specific operating procedures and disciplines during the course of evolution. Operating disciplines are allowed to change and evolve to better suit the search effort. The two major benefits anticipated are performance and robustness. Performance is measured by the goodness of the solution and by the computational effort required to obtain the solution. Robustness is related to the performance of the methodology for a wide range of optimization problems. A robust methodology is one which is not adversely affected by a prespecified operating discipline parameter. In this perspective, robustness is related to the generality of the methodology: its ability to perform well even when the search effort starts in a state not suitable for the specific conditions of the problem at hand. The methodology is demonstrated and evaluated by implementing a known-to-be-difficult class of scheduling problems, the single-machine deterministic scheduling problems in which the objective is to minimize both the earliness and the tardiness.

Computer control system for a metal cutting machine

Abstract This paper reviews the work carried out by the Industrial Research Laboratory in cooperation with the Machine Tool Laboratory at the University of Florida to design and build a Supervision System for Machining Centers. The objective of the Supervision System is to perform three main functions: Chatter Detection and Control, Cutter Breakage Detection, and Adaptive Feed Control [1], [2]. Our contribution to this project consists of providing low-level intelligence to the machine tool controller and high-level user-friendly operator interface (monitoring screen), so that the machine tool, as a unit will be more robust and efficient, thus more economical to operate. A simple PC based embedded control system for chatter and tool breakage detection as well as a user-friendly operator interface unit will be discussed.

A new standard for industrial control languages

Abstract There exists a need and an opportunity to develop industrial control languages that take advantage of the powerful facilities offered by modern microcontrollers. Developing a fundamental systems description is a prerequisite to writing such high-level control languages. Such a fundamental systems description requires building and experimenting with alternative hardware/software architectures, and gathering empirical data in actual manufacturing environments. A 3-year project is initiated in the Industrial Research Laboratory at the University of Florida to investigate the various aspects of controller architectures, operating systems and languages. This report summarizes our approach and outlines the progress to date.

Symbolic Solution of Kronecker-Based Structured Markovian Models

This paper describes a method to obtain symbolic solution of large stochastic models using Gauss-Jordan elimination. Such solution is an efficient alternative to standard simulations and it allows fast and exact solution of very large and complex models that are hard to be dealt even with iterative numerical methods. The proposed method assumes the system described as a structured (modular) Markovian system with discrete states for each system module and transitions among those states ruled by Markovian processes. The mathematical representation of such system is made by a Kronecker (Tensor) formula, i.e., a tensor formulation of small matrices representing each system module transitions and occasional dependencies among modules. Preliminary results of the proposed solution indicate the expected efficiency of the proposed solution.

A general theory on spectral properties of state-homogeneous finite-state quasi-birth–death processes

Abstract In this paper a spectral theory pertaining to Quasi-Birth–Death Processes (QBDs) is presented. The QBD, which is a generalization of the birth–death process, is a powerful tool that can be utilized in modeling many stochastic phenomena. Our theory is based on the application of a matrix polynomial method to obtain the steady-state probabilities in state-homogeneous finite-state QBDs. The method is based on finding the eigenvalue–eigenvector pairs that solve a matrix polynomial equation. Since the computational effort in the solution procedure is independent of the cardinality of the counting set, it has an immediate advantage over other solution procedures. We present and prove different properties relating the quantities that arise in the solution procedure. By also compiling and formalizing the previously known properties, we present a formal unified theory on the spectral properties of QBDs, which furnishes a formal framework to embody much of the previous work. This framework carries the prospect of furthering our understanding of the behavior the modeled systems manifest.