Elsayed A. Orady

Integrated diagnosis using information-gain-weighted radial basis function neural networks

A new approach, the information-gain-weighted radial basis function neural network (RBFNN), has been proposed for machinery diagnosis in a manufacturing environment. This approach is based on the composite neural network, in which a series of RBFNNs are integrated together to perform the task of classification. Each RBFNN has only one output node and is treated as a sub-network. Unlike the conventional RBFNNs, in which only the outputs of hidden nodes are weighted, the information-gain-weighted RBFNN uses a weighting vector also in its input layer. In addition, the weighting vector is obtained according to the information gain of each input index to the process of diagnosis. In the diagnosis strategy, one sub-net is responsible for diagnosing one specific fault, while the composite network as a whole can diagnose different kinds of faults. By this approach, classification among known faults can be made and a novel fault, if any, can also be identified. This approach has been tested in the diagnosis of an internal thread tapping process. The results showed that the information-gain-weighted RBFNN can produce better distinction between conditions and the scheme of a composite neural network is indeed an improved structure for machinery diagnosis in the manufacturing environment.

Virtual reality software for robotics and manufacturing cell simulation

Virtual reality software (VRS) for robotics and manufacturing cells simulation is a tool to obtain a 3D graphics models for the manufacturing lines. A 3D graphics model will help facility planners to visualize the system before constructing it, make alternative designs, program robot paths, obtain layouts for the systems, obtain data for the discrete event simulation, and develop the cell control program. This paper presents an overview of VRS software, a proposal for the integration of VRS and discrete event simulation, and case study for manufacturing cell development.

Evaluation of Minimum Zone Straightness by a Nonlinear Optimization Method

In this paper, a new algorithm, based on a nonlinear optimization method (NOM), has been developed. The accuracy as well as the reliability/robustness of the new algorithm have been verified by applying it to more than 200 CMM measured data sets on differently manufactured parts. The results have been compared with that of Least Squares Method (LSM) and Convex Hull (CVH) method applied to the same data sets. A data filter is proposed to be enclosed in the new algorithm to detect and delete outliers in the data sets.

Drilling characteristics of an E-glass fabric-reinforced polypropylene composite and an aluminum alloy: A comparative study

This paper presents an experimental study on the drilling characteristics of an E-glass fabric-reinforced polypropylene composite and aluminum alloy 6061-T6. Both materials have many similar structural applications, especially in the automotive industry. The drilling characteristics considered were axial thrust force, torque, temperature increase during drilling, and chip morphology. Both axial thrust force and torque were significantly higher for the aluminum alloy but were independent of the cutting speed for both materials. However, both increased linearly with increasing feed rate for the composite, but nonlinearly for the aluminum alloy. The Shaw-Oxford equation for predicting axial thrust and torque worked well with the aluminum alloy but did not fit the composite’s axial thrust and torque characteristics. Both materials exhibited temperature rise at locations close to the drilled hole. The temperature rise decreased with increasing feed rate as well as increasing cutting speed; however, the maximum temperature rise in the composite was significantly lower than that in the aluminum alloy. There was also a significant difference in the morphology of chips of these two materials.

A knowledge-based system for fault diagnosis and maintenance of advanced automated systems

Abstract A knowledge-based system for diagnosis & maintenance of robotic systems has been designed which may also serve as a training tool for maintenance personnel. The system uses a Group Technology (GT) approach for fault classification and analysis. A simple user interface leads the user through a consultation phase in order to arrive at a diagnosis. It then recommends corrective actions and a simple test procedure to verify that the robot will perform satisfactorily after the repairs.

Genetic algorithm and fuzzy C-means based multi-voting classification scheme in data mining

This paper presents a practical scheme used in data mining for classifications based on fuzzy logic and multivoting decision algorithms. It combines the information gain heuristic and genetic algorithm (GA) to minimize the uncertainty level when estimating the weighting functions used in the multiple voting decision scheme. A preliminary test of this scheme using a well-know data set demonstrated its competency and performance improvement for classifications.

Capability study of robotics and manufacturing cell simulation software

In this paper, a capability analysis was developed for comparing Virtual Reality Software (VRS) of robotics and manufacturing cells simulation. The capability analysis included all basic features that should be maintained in a software for simulation of a large scale agile manufacturing system. An algorithm has been developed for this analysis which requires subdividing the attributes into business attribute, cost attributes and other attributes related to automotive industry.

Minimum Zone Evaluation of Cylindricity Using Nonlinear Optimization Method

Minimum Zone Evaluation (MZE) of cylindricity, as mathematically defined in ASME standard, yields a nonlinear optimization problem which could be an arduous task due to the convergence difficulty. In this paper, the problem is formulated by introducing a new expression and simplification. The contribution of the new formulation to the problem implementation is revealed through the investigation of Objective Function Images (OFIs). This formulation possesses improved characteristics in comparison with the conventional one. This improvement is one of the keys to a successful optimum procedure. Based on the new formulation and the simplex searching technique, an algorithm has been accomplished and verified statistically and experimentally. The verification results are satisfactory in view of accuracy and robustness of the algorithm.

Thermal EMF method for monitoring drilling tool wear

This paper describes a techniQue for on-line monitoring of drilling tool wear based on the thermal EMF (electromotive force) signal. The EMF signal was obtained from a natural thermocouple consisting of the tool (H. S. S.) and workpiece (AISI 1045) metals. The natural thermocouple is thus used as a kind of functional sensor which is sensitive to the cutting zone. The signal was collected and analyzed for three experiments at different cutting conditions. Analysis was carried out in time, frequency and amplitude domains. Several indices for the EMF signal were computed and their relationships with the tool wear were constructed. The results showed that the thermal EMF signal can be used to identify the occurrence of abnormal tool wear on major cutting edges and can indicate the end of tool life. Tool breakage can also be predicted. Consequently, a methodology for monitoring drilling tool wear can be established.

An automatic steplength control algorithm for stiff ODEs systems

processes when analyzing their thermal behavior. The finite element formulation for such a problem on a process domain produces a system of ordinary differerential equations which is characterized to be ‘stiff. Careful handling of the solution is required in order to avoid oscillatory and inaccurate results. This problem can be avoided by selecting a suitable recurrence scheme together with an efficient steplength control algorithm. In this paper an automatic steplength control algorithm has been developed and implemented in a finite element program for the transient convective-transport equation. The algorithm has been tested using an example representing a manufacturing process. The conv~tiv~trans~rt equation is applied to many manufacturing processes to determine the transient temperature distribution [l-7]. It is applied to metal cutting, metal casting, forming, welding, etc. The analytical solution is found to be difficult and often not possible in such applications due to the complexity of the problem [4]. Therefore, numerical methods such as finite difference and finite element are usually employed. The finite element fo~ulation produces a system of first-order ordinary differential equations (ODES). The system of ODES is found to be, in most cases, very ‘stiff’ and produces oscillatory results when the time step is inadequate at the initial stages of the solution [8,9]. In order to avoid these problems, there is a need for an automatic time steplength control algorithm which can be used to automatically determine the time steps for a stable, accurate and economic solution [IO]. This paper presents a new automatic steplength control algorithm that is used with suitable recurrence schemes in solving such systems of ODES. It also gives a discussion for the formulation of the problem, and the selection of the recurrence schemes.