No description available.

<p>The impinging jets atomizer is widely used in engineering applications. As two liquid jetsimpinging to each other, a liquid sheet is first formed and then breaks up into small dropletsto comply the atomization. The features such as size, shape, velocity, thickness, etc., of thesheet/droplet are controlled by various impingement parameters such as impinging angle,jet velocity, and physical properties of the liquid. Since the sheet generation is prior to thedroplet, the modeling of the sheet is the premise of the droplet modeling. Therefore, to studythe atomization of the impinging jet atomizer, it is important to pay effort on the research ofimpinging sheet both experimentally and theoretically. In this research, the characterizationof the impinging sheet formed by two jets is given in two specific aspects, the thicknessand the velocity. A non-intrusive measurement technique, partial coherent interferometry(PCI) is developed and applied to measure the thickness of the impinging sheet dynamically.The PCI unitizes the calibrated linear relationship between the optical path difference andthe degree of coherence to measure the impinging sheet thickness. By placing the sheet inone of the two branches of the designed interferometer, the optical path is altered basedon the sheet thickness and shown as the change of the degree of coherence of interferencepattern recorded by the camera. With a calibration process, the thickness of the sheet is thencan be measured via a designed interferometer. The velocity measurement of the impingingsheet is implemented via particle tracking velocimetry (PTV) adopted with the shadowgraphtechnique. To implement the particle tracking velocimetry, seeding particles are added intothe fluid and with the aid of an imaging acquiring system and the post-processing algorithm,the locations of those particles in different frames are identified. Thus, the velocity of the fluidis estimated as the velocity of the particles calculated from the recorded images. However,while applying the PTV to investigate the impinging sheet studied in this research, theparticles can be recorded at a large field of view with insufficient magnification. This is ownedto the so-called "particle induced lens effect" found when applying the small particles to athin liquid sheet. When the seeding particles move to the region where the sheet thicknesshas a similar scale as the particle, the fluid will wrap around the particle and act as a positivelens. For shadowgraph imaging, the collimated light forms an enlarged shadow at the image plane by passing through such lens. Experimentally, the thickness measurements via PCIare implied to the impinging sheet generated under a range of Reynolds number between 269to 370 and velocity measurements via PTV are implied to the ones under Reynolds numberof 362 to 430. The measured results for both thickness and velocity are different from thetheoretical model of the impinging sheet which implies the need for a review of sheet model.Therefore, in this research, the author proposed a revised impinging sheet model considerthe friction effect due to the air over the sheet. A theoretical analysis is made base on theboundary layer equation under the cylindrical coordinate with unique boundary conditionsassumed for the impinging sheet. By introducing the unique similarity variable found byauthor, the equation could be transformed to an ordinary differential equation and solvednumerically. The revised model first predict the air boundary layer profile over the sheet,then, estimate the sheet velocity profile as a function of the distance to the impinging pointand the azimuth angle. As a parameter of the revised sheet model, the jet velocity profilebefore the impingement is also assumed as a free jet gradually developed from a Posieuilleflow and estimated in advance. The revised model is compared with the experimental resultsand some key parameters are identified empirically.</p> <p>Other than the thickness and velocity, this research is also interested in measuring thegeometry of the sheet and the detached droplets. Thus, a multi-view digital inline holography(DIH) technique is developed to capture the three-dimensional shape of the impinging sheetand the locations of the droplets. The DIH determines the shape and location of the targetin a detection volume base on the recorded hologram. The MvDIH, as the name suggested,combines the DIH results from multiple orientations to reconstruct the shape and the locationof the target. Two reconstruction ideologies, cross-section based one and the outline basedone, are proposed. The former estimates the target by finding the intersection of the recordedcross-sections of the target from different views. The latter estimates the target geometryby combining the outlines determined by DIH at different views. To evaluate the feasibilityof such technique, a test model which imitates the droplet and liquid ligament structure isapplied to the measurement in this research. Yet, the application on a real impinging sheetis not implemented.</p>

A novel control method called orthogonal eigenstructure control is developed for active vibration cancellation in structures. Orthogonal eigenstructure control is a feedback control method applicable to multi-input multi-output linear systems. While the available control design methodologies offer a large and complex design space of options that can often overwhelm a designer, this control method offers a significant simplification of the design task while still allowing some experience-based design freedom. For example, eigenstructure assignment methods need definition of a desired eigenvector for the closed-loop system. The controller designer may also be required to do pole placement. Considering the fact that there are no one-to-one relationships between the elements of the closed-loop eigenvectors of a model and the states of the system, this effort could be inefficient for many practical systems. Moreover, for large-scale systems, defining or shaping the eigenstructures become a relatively difficult task. Orthogonal eigenstructure control is a state feedback-like control law that is relatively easy to design and implement to multiple-input multiple-output systems. It allows control engineers to achieve good performing designs even with little design experience, while the existing methods are highly dependent on designer experience. Orthogonal eigenstructure control is introduced and extended to be applicable to linear systems regardless of the number and location of the actuators and sensors. Also, the concept of progressive application of the proposed control method for increasing robustness is described. Finally, the result of application of the control method for vibration cancellation of a test plate is investigated through experiments for tonal and wideband disturbances. The results show a significant reduction of vibrations using the orthogonal eigenstructure control with relative ease in finding the control gain matrix. / Ph. D.

A robust nonlinear observer, utilizing the sliding mode concept, is developed for the dynamic positioning of ships. The observer provides the estimates of linear velocities of the ship and bias from the slowly varying environmental loads. It also filters out wave frequency motion to avoid wear of actuators and excessive fuel consumption. Especially, the observer structure with a saturation function makes the proposed observer robust against neglected nonlinearties, disturbances and uncertainties. A direct adaptive neural network controller is developed for a model of an underwater vehicle. Radial basis neural network and multilayer neural network are used in the closed-loop to approximate the nonlinear vehicle dynamics. No prior off-line training phase and no explicit knowledge of the structure of the plant are required, and this scheme exploits the advantages of both neural network control and adaptive control. A control law and a stable on-line adaptive law are derived using the Lyapunov theory, and the convergence of the tracking error to zero and the boundedness of signals are guaranteed. Comparison of the results with different neural network architectures is made, and performance of the controller is demonstrated by computer simulations. The sliding mode observer is used to eliminate observation spillovers in the vibration control of flexible structures. It is common to build a state feedback controller and a state estimator based on the mathematical model of the system with a finite number of vibration modes, but this may cause control and observation spillover due to the residual (uncontrolled) modes. The performance of a sliding mode observer is compared with that of a conventional Kalman filter in order to demonstrate robustness and disturbance decoupling characteristics. Simulation and experimental results using the sliding mode observer are presented for the active vibration control of a cantilever beam using smart materials. / Ph. D.

We consider geometric and material nonlinearities when studying numerically, by the finite element method, transient three-dimensional electroelastic deformations of a graphite-epoxy square plate sandwiched between two piezoceramic (PZT) layers. Points on the four edges of the bottom surface of the plate are restrained from moving vertically. The two opposite edges of the plate are loaded by equal in-plane compressive loads that increase linearly with time and the other two edges are kept traction free. The plate material is modeled as orthotropic and neoHookean. For the transversely isotropic PZT the second Piola-Kirchhoff stress tensor and the electric displacement are expressed as second degree polynomials in the Green-St. Venant strain tensor and the electric field. Both direct and converse piezoelectric effects are accounted for in the PZT. The plate is taken to have buckled when its centroidal deflection equals three times the plate thickness. The dynamic buckling load for the plate is found to strongly depend upon the rate of rise of the applied tractions. With the maximum electric field limited to 1kV/mm, the buckling load is enhanced by 18.3% when the PZT elements are activated. For a peak electric field of 30kV/mm, the buckling load increased by 58.5%. When more than 60% of the surface area of the top and the bottom surfaces of the plate are covered by the PZT layers, then square PZT elements placed symmetrically about the plate centroid provide a larger enhancement in the buckling load than rectangular shaped or cross-shaped PZT elements. An increase in the plate thickness relative to that of the PZT actuators decreases the effectiveness of the PZT in enhancing the buckling load for the plate. The finite element code was modified to also analyze, in time domain, transient deformations of a viscoelastic material for which the second Piola-Kirchhoff stress tensor is expressed as a linear functional of the strain history of the Green-St. Venant strain tensor. It was used to analyze three-dimensional deformations of a thick laminated plate with layers made of aluminum, a viscoelastic material and a PZT. The following two arrangements of layers are considered. In one case a central PZT layer is surrounded on both sides by viscoelastic layers and aluminum layers are on the outside surfaces. The PZT is poled in the longitudinal direction and an electric field is applied in the thickness direction. Thus shearing deformations of the PZT layer are dominant. In the second arrangement, the aluminum layer is in the middle and the PZT layers are on the outside. The poling direction and the electric field are in the thickness direction; thus its extensional deformations are predominant. Three indices are used to gauge the damping of motion of plate particles, and the effectiveness of PZT actuators in enhancing this damping. It is found that the optimum thickness of the viscoelastic layers for maximum total energy dissipation is the same for each set-up. Also, the total thickness of the PZT layers which results in the maximum value of one of these indices of energy dissipation is the same for the two set-ups. Both arrangements give the largest value of this index for a plate of aspect ratio 10. Buckling behavior of a sandwich plate containing a soft core is also studied. The effects of the ratio of the elastic moduli of the outer layers to those of the core, and of the core thickness on the buckling load are analyzed. The top and the bottom layers are connected by very stiff blocks on two opposite edges where in-plane compressive time-dependent tractions are applied. / Ph. D.

This thesis presents the development and testing of distributed vibration absorber designs with specific application to heavy plates for the reduction of vibration and sound radiation. Two particular designs, already under investigation for use on thin panels or composite materials, were adapted to passively reduce broadband vibration and noise from large and heavy plates. These absorbers are referred to as Distributed Vibration Absorbers [DVAs] and Heterogeneous [HG] Blankets. Numerical models were developed, based on the theory of sound propagation through layered media and the vibration of plates, to simulate the performance of such absorbers for a variety of applications and media characteristics. The new absorber designs were then tested on a large, marine-type plate (4 feet by 2 feet by 1/4 inch) and showed both broadband noise and vibration control from 60 Hz to 5 kHz. DVAs could reduce the vibrating plate resonance magnitudes on the order of 15 dB at their tuning frequencies while providing overall vibration reduction of 5 dB or greater at higher frequencies. HG blankets were also capable of reducing plate resonance vibration up to 15 dB at their tuning frequencies and produced overall vibration reduction of 5 dB at higher frequencies. These absorbers are entirely passive, i.e. requiring no controller or prior modal testing of the structure, were placed randomly during testing, and are designed to contribute less than 10% additional mass to the structure, making them a robust vibration and noise control solution. / Master of Science

Instead of dissipating vibration energy into heat waste via viscous damping elements, this dissertation proposes an innovative vibration control method which can simultaneously mitigate vibration and harvest the associated vibration energy using electromagnetic energy harvesters. This dissertation shows that the electromagnetic energy harvester can work as a controllable damper as well as an energy harvester. The semi-active control of a linear electromagnetic energy harvester, for improvement of suspension performance, has been experimentally implemented in a scaled-down quarter-car suspension system. While improving performance, power produced by the harvester can be harvested through energy harvesting circuits. This dissertation also proposes a mechanical-motion-rectifier(MMR)-based electromagnetic energy harvester using a ball-screw mechanism and two one-way clutches for the application of replacing the viscous damper in vehicle suspensions. Compared to commercial linear harvesters, the proposed design is able to provide large damping forces and increase power-dissipation density, making it suitable to vehicle suspensions. In addition, the proposed MMR-based harvester can convert reciprocating vibration into unidirectional rotation of the generator. This feature significantly increases energy-harvesting efficiency by enabling the generator to rotate at a relatively steady speed during irregular vibrations and improves the system reliability by reducing impact forces among transmission gears. Extensive theoretical and experimental analysis have been conducted to characterize the proposed MMR-based energy harvester. The coupled dynamics of the suspension system with the MMR-based energy harvester are also explored and optimized. Furthermore, a new control algorithm is proposed to control the MMR-based energy harvester considering its unique dynamics induced by the one-way clutches. The results show that the controlled proposed electromagnetic energy harvester can possibly improve ride comfort of vehicles over conventional oil dampers and simultaneously harvest the associated vibration energy. / Ph. D.

Traditional dynamic vibration absorbers (DVAs) consist of a mass-spring-damper system and are an effective means of attenuating structural vibration over a narrow frequency band. The effective bandwidth of the DVA can be increased by the addition of an externally controlled force, generally applied between the reaction-mass and the primary structure. Such devices are known as hybrid DVAs. This thesis presents a new hybrid DVA configuration which utilizes two reaction-masses in parallel. On this proposed hybrid dual-mass (DM) DVA, the control force is applied between the reaction-masses. It is shown that in broadband control applications, the proposed DM-DVA requires less control force to achieve the same primary attenuation as the traditional hybrid single-mass (SM) DVA. The hybrid DM-DVA was compared to the hybrid SM-DVA with two tests. A numerical simulation of the hybrid DVAs attenuating a single-degree-of-freedom structure was performed. To achieve an equal amount of primary attenuation, the hybrid SM-DVA required 65% higher root-mean-square (RMS) control effort than the hybrid DV-DVA. The numerical model also demonstrated that the hybrid DM-DVA was less sensitive to changes in the system as compared to the hybrid SM-DVA. Additionally, a prototype hybrid DVA was built which could be configured as either the hybrid SM or DM-DVA. The prototype hybrid DVA was used with the feedforward Filtered-X LMS algorithm to control the vibration of a fixed-free beam. The hybrid SM and DM-DVAs attenuated the primary response by a factor of 11.5 and 12.3, while requiring control efforts of 4.9 and 2.7 V/N RMS, respectively. Thus, the hybrid DM-DVA required 45% less control effort while yielding a higher attenuation ratio in this experiment. These results demonstrate the superior performance of the proposed DM-DVA for broadband control applications as compared to the traditional SM-DVA. / Master of Science

Passive control techniques to minimize structural vibrations are limited with respect to the amount of attenuation obtained especially in the low-frequency region but do not require adding any power. Active control methods are effective for reducing structural vibrations, especially at low frequencies, but may require significant control effort. Thus, passive and active control methods have complementary frequency ranges of application. This research consists of combining active and passive control techniques to simultaneously attenuate extensional and flexural power flows in infinite thin beams and determine the advantages and disadvantages of such a combination. An analytical model is developed for an infinite beam with a passive insert of high damping placed at some distance from a point force excitation (passive approach). The passive control of vibrations results in a reduction of both extensional and flexural power flows downstream of the passive material discontinuity. The simultaneous active control of extensional and flexural waves, using two co-located independent piezoceramic actuators bonded to the surface of the beam, is theoretically studied. The active control model shows that the use of two independent piezoceramic actuators allows complete cancellation of the total power flow (sum of the extensional and flexural power flows) downstream of the actuators. The combination of passive and active control methods for three different configurations (actuators located upstream of, downstream of, and on the passive insert) is investigated and complete control of the total power flow is again achieved. The results demonstrate that in the case of the actuators bonded to the passive material discontinuity, the active/passive combination has great potential for reducing the control effort required for the active controller. Finally, an approximation of the influence of heavy fluid flanking paths on the optimal active/passive system is developed by simulation of these flanking paths using axial and torsional springs. This last study shows that both axial and torsional springs will result in modification of the control effort required by the actuators if their respective stiffness is greater than the equivalent stiffness of the section in parallel with the springs. / Master of Science

No / This paper presents an investigation into the impact of data dependencies in real-time high performance sequential and parallel processing. An adaptive active vibration control algorithm is considered to demonstrate the impact of data dependencies in real-time computing. The algorithm is analysed in detail to explore the inherent data dependencies. To minimize the impact of data dependencies, an investigation into reducing memory access in sequential computing is provided. The impact of data dependencies with various interconnections is also explored and demonstrated in real-time parallel processing through a set of experiments.