College of Engineering and Computer Science

An on-line scheme for monitoring tool wear in unmanned machining operations using artificial neural networks (ANNs) is proposed. Various configurations of ANNs are studied to increase the accuracy of tool wear estimation. With this aim three configurations of the ANNs namely, an ANN without memory, an ANN with one phase memory, and an ANN with two phase memory are considered. Each ANN is trained to associate an input vector which consists of values of cutting conditions, with an output vector containing flank wear as a single output. The training data and evaluation data is generated using the popular analytical tool wear model. The performance of all the ANNs are compared by considering four different cases of evaluation data. The proposed scheme of tool wear modeling using ANNs is easily extendible to include other cutting parameters and can be implemented in real-time.

A methodology is presented to construct an approximate fuzzy-mapping algorithm that maps multiple inputs to single outputs given a finite training set of argument vectors functionally linked to corresponding scalar outputs. Its scope is limited to problems where the features are known in advance, or equivalently, where the expected functional representation is known to depend exclusively on the known selected variables. Programming and simulations to implement the methodology make use of Matlab Fuzzy and Neural toolboxes and a PC application of Prolog, and applications range from approximate representations of the direct kinematics of parallel manipulators to fuzzy controllers.

This dissertation presents a new type of two-dimensional, real time, color ultrasonic scanner able to measure and display brain metabolism by monitoring amplitudes of localized changes of intracranial interfaces. A real time image is obtained with a flexible array of transducers which eliminates the strong reflection from the skull due to a shape mismatch, and reduces the topological mislocations in the image. The image is generated by a superposition of a gray scale image representing static structures, and a color coded pattern representing motion information. The new technique of motion detection based on image subtraction features high accuracy and gives the scanner the unique capability to detect multidirectional motion of the intracranial interfaces, and to display the amplitude of the motion in real time. A series of experiments performed with the scanner demonstrates outstanding agreement between theoretical design and hardware performance. The scanner has been measured to have a lateral resolution of 4 mm, a temporal resolution of 30 fr/s, a motion detection resolution of 5 $\mu$m, a time gain compensation of 40 dB, and a signal/noise ratio of more than 40 dB. Successful tests, performed on a live human brain, show a well defined echo pattern arising from intracranial structures within the brain, and a strong correlation between the detected surface pulsations and heart beat has been observed. Monitoring the image synchronously with the heart beat and the external stimulus presence gives clinicians the unique opportunity of visualization of detailed cross-sectional anatomy of portions of the human brain, permitting direct observation, mapping the structure and function in a normal human brain, and studying the pathophysiology of mental illness by demonstrating structural metabolic, and neurochemical abnormalities. Additional experiments, performed on other parts of the human body, demonstrated clearly the advanced features of the scanner and its successful application to other areas of medicine beyond neurology. Furthermore, this new real time, two-dimensional brain scanner will be suitable for remote diagnosis and consultation, and long-distance delivery of quality health care via teleconferencing and telecommunication equipment. The modular design of the scanner allows blocks, such as multidirectional motion detection, and the flexible transducer array to be used as stand alone units or to be built into already existing ultrasonic equipment such as sonars, motion sensors, nondestructive testing of materials, etc.

High energy density PAN-based carbon fiber anode materials for lithium-ion type batteries were developed. Commercially available organic precursors were thermally converted to carbons. The effects of precursor material, carbonization temperature, heating ramp rate, soak time and gaseous atmosphere during the thermal treatment on the electrochemical performance of the carbon fibers were studied. In order to evaluate the electrochemical performance of the carbon fibers, test cells were assemble using the carbon materials prepared in the laboratory and intercalation/deintercalation experiments were performed. The results indicated that the highest reversible capacity and lowest irreversible capacity loss was obtained for carbon fibers carbonized at 1100C at fast ramp rate of 26C/min. X-ray diffraction experiments revealed a relation between the capacity and the irreversible capacity loss on first cycle, and the size of the crystallites Lc. A phenomenological explanation for this behavior was developed. Using electrochemical impedance spectroscopy the diffusion coefficient of Li in the tested carbon fibers was calculated. In addition, the influence of electrolyte composition (solvent and salt) on the reversible and irreversible capacities as well as on the intercalation/deintercalation potential profile was investigated. An electrolyte containing 1M LiPF6 in EC:DEC:DMC (40:30:30 v/o) proved to be most suitable for these carbon fiber materials improving significantly their electrochemical performance. Finally, coin cells were assembled containing the carbon fiber material prepared in the laboratory. They were tested for reversible and irreversible capacity. The coin cells proved that the synthesized carbon anode materials possess high energy density and could be used in commercial applications.

Catastrophic events in the past revealed the need for more research in the field of emergency evacuation. During such a procedure, different problems such as congestion at the related traffic networks because of the large number of the evacuating vehicles can occur. Current best practices, in order to deal with such problems, suggest the further involvement of buses in evacuation operations. On the first part of this study after the accurate development of the related simulation model, the optimization of a selected bus system characteristics focusing on the vehicle routing parameter will follow through the development and the application of a non-linear cost minimization problem. On the second part, the potential use of the regular-everyday bus routes in a no-notice emergency evacuation in order to save time comparing to the time needed so as to assign the actual evacuation routes to the evacuation bus vehicles will be analyzed.

Publish/subscribe is a powerful paradigm for distributed applications based on decoupled clients of information. In pub/sub applications, there exist a large amount of publishers and subscribes ranging from hundreds to millions. Publish/subscribe systems need to disseminate numerous events through a network of brokers. Due to limited resources of brokers, there may be lots of events that cannot be handled in time which in turn causes overload problem. Here arises the need of admission control mechanism to provide guaranteed services in publish/subscribe systems. Our approach gives the solution to this overload problem in the network of brokers by limiting the incoming subscriptions by certain criteria. The criteria are the factors like resources which include bandwidth, CPU, memory (in broker network), resource requirements by the subscription.

This thesis proposes to estimate the speed of a moving acoustic source by either linear or non linear processing of the resulting Doppler shift present in a high-frequency pilot tone. The source is an acoustic modem (Hermes) which currently uses moving average to estimate and compensate for Doppler shift. A new auto regressive approach to Doppler estimation (labeled IIR method in the text) promises to give a better estimate. The results for a simulated peak velocity of 2 m/s in the presence of additive noise showed an RMSE of 0.23 m/s using moving average vs. 0.00018 m/s for the auto regressive approach. The SNR was 75 dB. The next objective was to compare the estimated Doppler velocity obtained using the two algorithms with the experimental values recorded in real time. The setup consisted of a receiver hydrophone attached to a towing carriage that moved with a known velocity with respect to a stationary acoustic source. The source transmitted 375 kHz pilot tone. The received pilot tone data were preprocessed using the two algorithms to estimate both Doppler shift and Doppler velocity. The accuracy of the algorithms was compared against the true velocity values of the carriage. The RMSE for a message from experiments conducted indoor for constant velocity of 0.4 m/s was 0.6055 m/s using moving average, 0.0780 m/s using auto regressive approach. The SNIR was 6.3 dB.

This manuscript predicts the behavior of a doubly reinforced concrete beam with a circular opening at its midspan by closely analyzing traditional beam theory and design. It then confirms these predictions with finite element modeling software while providing design suggestions. The analysis is limited to the tensile and compressive stresses and cracking behavior. The objectives are to determine the stress distribution around a circular opening that agrees with conventional beam theory. The beam behavior is examined from zero load to failure load. ANSYS is utilized in lieu of real world testing, and the appendix includes the finite element results for a beam including design recommendations. The results lay the foundation for a possible new design procedure of concrete beams with single or multiple circular openings. This research offers useful information that was unavailable previously. More research can be conducted to help designers to design lighter, more efficient concrete beams.

The Center for Ocean Energy Technology at Florida Atlantic University is developing an ocean energy turbine system to investigate the feasibility of harnessing Florida's Gulf Stream current kinetic energy and transforming it into a usable form. The turbine system has components which are prone to marine corrosion given the materials they are made of and to the harsh environment they will be exposed to. This study assumes a two-part system composed of a coating system acting as a barrier and sacrificial anode cathodic protection which polarizes the metal structures to a potential value where corrosion is significantly reduced. Several configurations (varying in anode quantity, size and location) were considered in order to cathodically protect the structures with various coating qualities (poor, good and excellent). These cases were modeled and simulated via Boundary Element Method software and analyzed so as to assess the most appropriate design.

This paper presents the comparison of shrinkage and corrosion characteristics of optimized hybrid Rice Husk Ash (RHA)/Fly Ash (FA)-modified Concrete, with those of normal concrete in the marine environment. Uses of both FA and RHA have numerous environmental benefits. Shrinkage performance was determined by subjecting the mixes to restrained shrinkage testing per ASTM C1581. The time to cracking of the specimens improved an average of 18% with the hybrid mixes. Corrosion testing of reinforced columns was performed in a simulated tidal cycle Marine Environment. Corrosion potential improved by as much as 35% for the mix with the highest FA/RHA replacement, and corrosion activity as measured with potentiostat equipment improved by an average of 34% . These results indicate a clear performance improvement of the modified concrete that is proportional to the percent replacement of cement.