Engineering Commons^™

Comparison Of Particle Swarm Optimization And Backpropagation As Training Algorithms For Neural Networks, Ganesh K. Venayagamoorthy, Venu Gopal Gudise

Electrical and Computer Engineering Faculty Research & Creative Works

Particle swarm optimization (PSO) motivated by the social behavior of organisms, is a step up to existing evolutionary algorithms for optimization of continuous nonlinear functions. Backpropagation (BP) is generally used for neural network training. Choosing a proper algorithm for training a neural network is very important. In this paper, a comparative study is made on the computational requirements of the PSO and BP as training algorithms for neural networks. Results are presented for a feedforward neural network learning a nonlinear function and these results show that the feedforward neural network weights converge faster with the PSO than with the BP …

Control Of Cascaded Multi-Level Inverters, Keith Corzine, Fang Zheng Peng, J. Wang, M. W. Wielebski

Electrical and Computer Engineering Faculty Research & Creative Works

A new type of multi-level inverter is introduced which is created by cascading two three-phase three-level inverters using the load connection. This new inverter can operate as a nine-level inverter and naturally splits the power conversion into a higher voltage lower-frequency inverter and a lower-voltage higher frequency inverter. This type of system presents particular advantages to naval ship propulsion systems which rely on high power quality, survivable drives. New control methods are described involving both joint and separate control of the individual three-level inverters. Simulation results demonstrate the effectiveness of both controls. A laboratory set-up at the Naval Surface Warfare …

Voltage Balancing Control Of Diode-Clamped Multilevel Rectifier/Inverter Systems, Zhiguo Pan, Keith Corzine, V. R. Stefanovic, J. M. Leuthen, Fang Zheng Peng

Electrical and Computer Engineering Faculty Research & Creative Works

This paper presents a new voltage balancing control for the diode-clamped multilevel rectifier/inverter system. A complete analysis of the voltage balance theory for a 5-level back-to-back system is given. The analysis is based on fundamental frequency switching control and then extended to pulse-width modulation. The method involves obtaining optimal switching angles; a process which is described in detail in this paper. The proposed control strategy regulates the DC bus voltage, balances the capacitors, and decreases the harmonic components of the voltage and current. Simulation and experimental results demonstrate the validity of the optimizing method and control theory.

Adaptive Neural Network Based Power System Stabilizer Design, Wenxin Liu, Ganesh K. Venayagamoorthy, Donald C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

Power system stabilizers (PSS) are used to generate supplementary control signals for the excitation system in order to damp the low frequency power system oscillations. To overcome the drawbacks of conventional PSS (CPSS), numerous techniques have been proposed in the literature. Based on the analysis of existing techniques, this paper presents an indirect adaptive neural network based power system stabilizer (IDNC) design. The proposed IDNC consists of a neuro-controller, which is used to generate a supplementary control signal to the excitation system, and a neuro-identifier, which is used to model the dynamics of the power system and to adapt the …

An Industrial Food Processing Plant Automation Using A Hybrid Of Pi And Fuzzy Logic Control, Ganesh K. Venayagamoorthy, D. Naidoo, P. Govender

Electrical and Computer Engineering Faculty Research & Creative Works

An industrial food processing plant consisting of nonlinear dynamics requires pressure and temperature controllers to be nonlinear. In this paper, nonlinear controllers are designed using fuzzy logic to augment the conventional pressure and temperature PI controllers for an industrial food processing plant. Simulation results are presented to show improvement in the plant response with the hybrid controllers.

Distributed Sensor Network For Civil Infrastructure Health Monitoring, Jose Humberto Mata Hernandez

Masters Theses

"This thesis focuses on the development of a robust, low power consumption and cost effective wireless data acquisition, system to be used for civil infrastructure health monitoring. The complete solution that has been developed for this application integrates micro-electromechanical sensing elements (e.g. PVDF sensor); a wireless data acquisition system based on a Bluetooth-enabled radio, a microcontroller with analog to digital conversion capabilities, a power source including a solar cell array for long term operation, and a weatherproof enclosure for on-site installation. The complementary part of the solution is a weatherproof On-site Remote Access Point (ORAP) which collects, stores, and transmits …

Novel Concepts For Monitoring The Stability Of Open Pit Mine Highwalls Using Radar Interferometry And Spectral Imaging, Jami Girard Dwyer

Masters Theses

"Ground control problems at surface mining operations can occur for a variety of reasons. The geologic setting, rock strengths, joint spacing, orientation, pore pressures, and many other factors contribute to slope instabilities that range from small rock falls to massive slides of material. While much of this ground movement can be predicted or controlled, each year many completely unexpected failures occur. This poses serious safety and economic issues for the mining industry, since any unanticipated movement of the ground increases the potential for significant property damage, injuries, and even death. In the United States alone, 70 people died in slope …

Dsp Processor Based Vision Processing And System Integration For Control Of Laser Aided Manufacturing Process (Lamp), Srinivas Musti

Masters Theses

"This thesis presents the research and development of a real-time vision system and integration of various subsystems to build the hybrid system for the Laser aided manufacturing process (LAMP), being developed at the University of Missouri--Rolla. A description of the real-time control architecture, showing the interaction of the real-time controller with various sensors and subsystems to monitor and regulate the process is presented.

Control of melt pool geometry is important for ensuring the final product quality since the bead size significantly influences the joint strength. However, during the deposition process, impurities and the plume present around the pool complicate the …

Comparison Of Anode And Electrolyte Support Configuration Of Single-Chamber Sofc, Piotr Jasinski, Toshio Suzuki, Zach Byars, Fatih Dogan, Harlan A. Anderson

Materials Science and Engineering Faculty Research & Creative Works

The single-chamber solid oxide fuel cell (SOFC) has recently received considerable attention. This type of fuel cell operates in the mixed reactant gas and therefore does not require separation of the fuel and oxidant gases. In this paper we report the anode- and electrolyte-supported configurations of the single chamber SOFC that we have studied. The system based on typical fuel cell composition has been examined: yttrium stabilized zirconia (YSZ) as an electrolyte, nickel/YSZ cermet as an anode, lanthanum strontium manganese oxide (LSM) as a cathode. The cells were prepared by tape-casting and screen-printing techniques. The cathode processing parameters have been …

Comparison Of 3-, 5-, And 6-Phase Machines For Automotive Charging Applications, B. Zhang, Hua Bai, Walter Eversman, Ronald J. Krefta, Duane J. Buening, Steven Pekarek, Gregory R. Holbrook

Mechanical and Aerospace Engineering Faculty Research & Creative Works

In this paper, the performance of a 5-phase claw-pole machine/converter system is compared with those of 3-phase and 6-phase systems of the same volume. It is shown that the acoustic output of the 5-phase machine compares favorably with that of the 6-phase machine and is much lower than that of the 3-phase machine over a majority of the operating range. In addition, the output current produced by the 5-phase machine Is, in general, 6-8% higher than those of the 3-phase and 6-phase machines. These results suggest that for maximizing power density while minimizing cost and acoustic noise, 5-phase machines provide …

A Single-Element Tuning Fork Piezoelectric Linear Actuator, J. Satonobu, K. Nakamura, S. Ueha, Daniel S. Stutts, James R. Friend

Mechanical and Aerospace Engineering Faculty Research & Creative Works

This paper describes the design of a piezoelectric tuning-fork, dual-mode motor. The motor uses a single multilayer piezoelectric element in combination with tuning fork and shearing motion to form an actuator using a single drive signal. Finite-element analysis was used in the design of the motor, and the process is described along with the selection of the device's materials and its performance. Swaging was used to mount the multilayer piezoelectric element within the stator. Prototypes of the 25-mm long bidirectional actuator achieved a maximum linear no-load speed of 16.5 cm/s, a maximum linear force of 1.86 N, and maximum efficiency …

Approximate Dynamic Programming Based Optimal Neurocontrol Synthesis Of A Chemical Reactor Process Using Proper Orthogonal Decomposition, Radhakant Padhi, S. N. Balakrishnan

Mechanical and Aerospace Engineering Faculty Research & Creative Works

The concept of approximate dynamic programming and adaptive critic neural network based optimal controller is extended in this study to include systems governed by partial differential equations. An optimal controller is synthesized for a dispersion type tubular chemical reactor, which is governed by two coupled nonlinear partial differential equations. It consists of three steps: First, empirical basis functions are designed using the "Proper Orthogonal Decomposition" technique and a low-order lumped parameter system to represent the infinite-dimensional system is obtained by carrying out a Galerkin projection. Second, approximate dynamic programming technique is applied in a discrete time framework, followed by the …

Hierarchical Optimal Control Of A Turning Process - Linearization Approach, Anand Dasgupta, B. Pandurangan, Robert G. Landers, S. N. Balakrishnan

Mechanical and Aerospace Engineering Faculty Research & Creative Works

Machining process control technologies are currently not well integrated into machine tool controllers and, thus, servomechanism dynamics are often ignored when designing and implementing process controllers. In this paper, a hierarchical controller is developed that simultaneously regulates the servomechanism positions and cutting forces in a lathing operation. The force process and servomechanism system are separated into high and low levels, respectively, in the hierarchy. The high level goal is to maintain a constant cutting force to maximize productivity while not violating a spindle power constraint. This goal is systematically propagated to the lower level and combined with the low level …

Modeling Fracture In Laminated Automotive Glazing Impacted By Spherical Featureless Headform, Lokeswarappa R. Dharani, S. R. Mettu, S. Zhao, S. D. Barbat, L. Chai

Mechanical and Aerospace Engineering Faculty Research & Creative Works

Laminated glass consisting of two soda lime glass plies adhered by a polyvinyl butyral interlayer (PVB) is used for automotive glazing. This paper describes the application of a dynamic, nonlinear finite element method to investigate the failure modes of a laminated glass subjected to low-velocity impact with a spherical head form. Crack type, crack location and crack initiation time are evaluated using the maximum principal stress and J-integral criterion. Failure occurred due to flexural stresses and not bearing stresses. the first crack always initiated at the center of the outer impacted ply and PVB interface, and later on the exterior …

Multidimensional, Time-Accurate Cfd Simulation Of Adsorption/Desorption In A Carbon Canister, Xingyan Bai, Kakkattukuzhy M. Isaac, D. Klein, R. Banerjee, J. Edson, W. Breig, L. Oliver

Mechanical and Aerospace Engineering Faculty Research & Creative Works

Computational fluid dynamics simulations of fixed-bed adiabatic adsorption/desorption processes are presented in this paper. Linear driving force model is used for heat and mass transfer rates. a two-dimensional cylindrical canister and three-dimensional automotive production canister geometry are used to study the adsorption/desorption processes of carbon dioxide in helium carrier gas on Norit B4 activated carbon. the two-dimensional results compare well with the results of Hwang et al. [1]. Computational results as breakthrough curve, adsorption amount and temperature profiles are provided. Results show that non-adiabatic model should be used to fully utilize the activated carbon bed capacity prior to breakthrough. Copyright …

Enhancement Of Output Electric Power In Switched Reluctance Generators, S. Dixon, Babak Fahimi

Electrical and Computer Engineering Faculty Research & Creative Works

This paper introduces a new switching strategy to increase overall productivity for a switched reluctance generator (SRG). This paper presents the idea of using an intermediate freewheeling mode to boost the magnetizing current without taking additional energy from the DC link. Once the optimal motional back-EMF is established, then both switches will be turned off. Although during the freewheeling process no electricity is generated, by proper timing of each region an overall enhancement in productivity of the SRG is achieved. to demonstrate the practicality of the proposed technique, simulation results are provided. These simulations take into account the switching losses …

Million City Traveling Salesman Problem Solution By Divide And Conquer Clustering With Adaptive Resonance Neural Networks, Samuel A. Mulder, Donald C. Wunsch

Electrical and Computer Engineering Faculty Research & Creative Works

The Traveling Salesman Problem (TSP) is a very hard optimization problem in the field of operations research. It has been shown to be NP-complete and is an often-used benchmark for new optimization techniques. One of the main challenges with this problem is that standard, non-AI heuristic approaches such as the Lin-Kernighan algorithm (LK) and the chained LK variant are currently very effective and in wide use for the common fully connected, Euclidean variant that is considered here. This paper presents an algorithm that uses adaptive resonance theory (ART) in combination with a variation of the Lin-Kernighan local optimization algorithm to …

Optimization Of An Automotive Generator Using Evolutionary Programming, A. Koenig, J. Williams, Steven Pekarek

Electrical and Computer Engineering Faculty Research & Creative Works

One of the primary design requirements of automotive generation systems is maximizing power density subject to the constraint of minimizing overall system cost. However, with progress made in reduction of automotive drive train noise, the torque ripple of the generator has been found to be a dominant noise source under idle conditions at high electrical load. Thus, an added design constraint is the minimization of torque ripple produced by the machine. in order to meet these objectives, numerical tools have been developed to automate the design process. a primary tool is an evolutionary algorithm that has been integrated within a …

Welcome To The Special Issue: The Best Of The Best, Donald C. Wunsch, Mike Hasselmo, De Liang Wang, Ganesh K. Venayagamoorthy

Electrical and Computer Engineering Faculty Research & Creative Works

No abstract provided.

Combining Evolving Neural Network Classifiers Using Bagging, Sunghwan Sohn, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

The performance of the neural network classifier significantly depends on its architecture and generalization. It is usual to find the proper architecture by trial and error. This is time consuming and may not always find the optimal network. For this reason, we apply genetic algorithms to the automatic generation of neural networks. Many researchers have provided that combining multiple classifiers improves generalization. One of the most effective combining methods is bagging. In bagging, training sets are selected by resampling from the original training set and classifiers trained with these sets are combined by voting. We implement the bagging technique into …

Cooperative Cleaning For Distributed Autonomous Robot Systems Using Fuzzy Cognitive Maps, H. Subramanian, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Cooperative Autonomous Cleaning is a simple challenge that can be implemented with the help of Fuzzy Cognitive Maps (FCM) by simulating the actual thinking process of the human. The human mind organizes its thoughts in priorities and this feature could be exploited well if a priori knowledge of the system exists. This technique has been attempted here for a DARS.

Web Personalization Using Neuro-Fuzzy Clustering Algorithms, Kartik Menon, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

Different users have different needs from the same web page and hence it is necessary to develop a system which understands the needs and demands of the users. Web server logs have abundant information about the nature of users accessing it. In this paper we discussed how to mine these web server logs for a given period of time using unsupervised and competitive learning algorithm like Kohonen''s self organizing maps (SOM) and interpreting those results using Unified distance Matrix (U-matrix). These algorithms help us in efficiently clustering users based on similar web access patterns and each cluster having users with …

Sequence Dependent Mobility Of Short Oligonucleotides In Free-Solution Electrophoresis, James D. Sterling, Eve F. Fabrizio, Ali Nadim

Business and Information Technology Faculty Research & Creative Works

No abstract provided.

Query-Based Learning For Aerospace Applications, Donald C. Wunsch, Emad W. Saad, J. J. Choi, J. L. Vian

Electrical and Computer Engineering Faculty Research & Creative Works

Models of real-world applications often include a large number of parameters with a wide dynamic range, which contributes to the difficulties of neural network training. Creating the training data set for such applications becomes costly, if not impossible. In order to overcome the challenge, one can employ an active learning technique known as query-based learning (QBL) to add performance-critical data to the training set during the learning phase, thereby efficiently improving the overall learning/generalization. The performance-critical data can be obtained using an inverse mapping called network inversion (discrete network inversion and continuous network inversion) followed by oracle query. This paper …

Smart Bridges With Fiber-Optic Sensors, Steve Eugene Watkins

Electrical and Computer Engineering Faculty Research & Creative Works

This article describes an approach to bridge improvements that uses smart structures, introduces key technologies for health monitoring systems based on fiber optics and gives an overview of two instrumented bridges. A sensing system in one application monitors general performance and health of the structure, while the system in a second application interrogates the behavior of a major structural repair.

Statistical Simulation Models For Rayleigh And Rician Fading, Chengshan Xiao, Y. Rosa Zheng, N. C. Beaulieu

Electrical and Computer Engineering Faculty Research & Creative Works

New simulation models are proposed for Rayleigh and Rician fading channels. First, the statistical properties of Clarke's fading model with a finite number of sinusoids are analyzed. An improved Clarke's model is then proposed for the simulation of Rayleigh fading channels. Based on this improved Rayleigh fading model, a novel simulation model is proposed for Rician fading channels. The new Rician fading model employs a zero-mean stochastic sinusoid as the specular (line-of-sight) component, in contrast to all existing Rician fading simulators that utilize a non-zero mean deterministic specular component. The statistical properties of the proposed Rician fading model are analyzed …

The Development Of A Closed-Form Expression For The Input Impedance Of Power-Return Plane Structures, Minjia Xu, Todd H. Hubing

Electrical and Computer Engineering Faculty Research & Creative Works

In multilayer printed circuit boards, the noise on the power bus is influenced by the impedance between the power and ground planes. Power-bus noise estimates require an accurate estimate of the power-bus input impedance. This paper develops a closed-form estimate of the input impedance for circular power-return plane structures. When the structure is lossy (e.g., boards employing embedded capacitance or densely populated boards), the energy reflected from the board edge does not significantly affect the input impedance. In general, the expressions developed here for circular structures can be used to estimate the impedance of lossy power-return plane structures of any …

The Subcircuit Extraction Problem, Nian Zhang, Donald C. Wunsch, Frank Harary

Electrical and Computer Engineering Faculty Research & Creative Works

The steady down scaling of CMOS device dimensions has been the main stimulus to the growth of microelectronics and computer aided very large scale integration (VLSI) design. But the more an integrated circuit (IC) is scaled, the higher its packing density becomes. The increasing size of chips, measured in either area or number of transistors, and the waste of the large capital investment involved in fabricating and testing circuits that do not work, make layout analysis and verification an important part of physical design automation. The most efficient way to overcome these difficulties is to identify a related collection of …

A Continually Online Trained Neurocontroller For The Series Branch Control Of The Upfc, Ganesh K. Venayagamoorthy, Radha P. Kalyani

Electrical and Computer Engineering Faculty Research & Creative Works

The crucial factor affecting the modern power systems today is load flow control. The Unified Power Flow Controller (UPFC) provides an effective means for controlling the power flow and improving the transient stability in a power network. The UPFC has fast complex dynamics and its conventional control is based on a linearized model of the power system. This paper presents the design of a neurocontroller that controls the power flow and regulates voltage along a transmission line. The continually online neurocontroller is used for controlling the series inverter of UPFC. Simulation results carried out in the PSCAD/EMTDC environment are presented …

A Discrete-Time Model For Spatio-Temporally Correlated Mimo Wssus Multipath Channels, Chengshan Xiao, Jingxian Wu, Sang-Yick Leong, Y. Rosa Zheng, Khaled Ben Letaief

Electrical and Computer Engineering Faculty Research & Creative Works

In this paper, a statistical discrete-time model is proposed for simulating wideband MIMO channels which experience spatially and temporally correlated, widesense stationary uncorrelated scattering (WSSUS) multipath Rayleigh fading. A new method is also presented to efficiently generate the correlated MIMO channel coefficients, which can be used for accurate simulation of physical continuous-time MIMO channel. The statistic accuracy of the discrete-time MIMO channel model is rigorously verified through theoretical analysis and extensive simulations in different criteria.