Taiwan is a small island state surrounded by the oceans but with large population. With limited land space, it would be worthwhile considering how to stabilize the existing coast or to create stable artificial beaches. Under the onslaught of storm surge and large wave from typhoons, beach erosion would occur accompanying by formation of a submerged bar beyond the surf zone with the sand removed from the beach. After the storm, the bar material maybe transport back by the swell and predominant waves which helps recover the original beach, thus producing a beach profile in dynamic equilibrium. The main purpose of this research is to use the back-propagation neural network¡]BPNN¡^, which trains a sample model and creates a system for the estimation, prediction, decision making and verification of an anticipated event. By the BPNN, we can simulate the key characteristic parameters for the storm beach profile resulting from typhoon action. Source data for training and verification are taken from the experimental results of beach profile change observed in large-scale wave tank¡]LWT¡^conducted by Coastal Engineering Research Center¡]CERC¡^in the USA in the 1960s and that from the Central Research Institute of Electric Power Industry in Japan in the 1980s. Some of the data are used as training pairs and others for verification and prediction of the key parameters of berm erosion and bar formation. Through literature review and simulation on the related parameters for storm beach profile, methodology for the prediction of the beach profile and bar/berm characteristics can be established.

Terpios hoshinota is coral-killing sponge which grows and covers most stony corals in shallow waters. It was first discovered at Green Island in 2006, and have since killed a lot of coral, yet we know little about the physiology of Terpios hoshinota. This research focuses on the propagation and growth of the sponge. In the sexual reproduction part, we collected tissue samples in 2009 and 2010, the sperm cells were found only in Jun and Aug in 2009. The oocytes were found in Apr, Jul, and Aug. In 2010, embryos occurred. No lunar pattern was found in a high-frequency sampling of tissues comparing the occurrence and sizes of oocytes and embryos. Embryos are more likely to be found in the central part of the sponge; this pattern does not apply to oocytes nor to sperm cells. The sponge may be hermaphroditic male and female gametes are developed at different locations or times. The sponge fragments can reinfect new host corals, although such capability decreased with increasing number of days suspending in the water column. The spicules parallel to each other and to the growth axis in tissue threads, moreover, the sponge quickly extended numerous tissue threads in the absence of adequate coral substrate, which may serve the function of reaching new hosts. The sponge grows faster under light than under dimmed conditions. Fusion of tissues could occur between non-identical genotypes, and allografting pairs of tissues have higher rates of rejection than isografting pairs. After allografting the sponge fragments from different areas, the fusion rates were depended on the distance of two populations in the northern coast of Green Island. The results supported that self-seeding is the mechanism how Terpios hoshinota populations exploded in the north coast Green Island. The ability to cross to the neighboring corals, to propagate by fragments, and to produce embryos may have all contributed to their self-seeding capability.

Wave propagation in unbounded media is a topic widely studied in different science and engineering fields. Global and local absorbing boundary conditions combined with the finite element method or the finite difference method are the usual numerical treatments. In this dissertation, an alternative is investigated based on the dynamic stiffness and the exponential window method in the space-wave number domain. Applying the exponential window in the space-wave number domain is equivalent to introducing fictitious damping into the system. The Discrete Fourier Transform employed in the dynamic stiffness can be properly performed in a damped system. An open boundary in space is thus created. Since the equation is solved by the finite difference formula in the time domain, this approach is in the time-wave number domain, which provides a complement for the original dynamic stiffness method, which is in the frequency-wave number domain. The approach is tested through different elasto-dynamic models that cover one-, two- and three-dimensional problems. The results from the proposed approach are compared with those from either analytical solutions or the finite element method. The comparison demonstrates the effectiveness of the approach. The incident waves can be efficiently absorbed regardless of incident angles and frequency contents. The approach proposed in this dissertation can be widely applied to the dynamics of railways, dams, tunnels, building and machine foundations, layered soil and composite materials.

Turbidite reservoirs in deep-water depositional systems, such as the oil fields in the offshore Gulf of Mexico and North Sea, are becoming an important exploration target in the petroleum industry. Accurate seismic reservoir characterization, however, is complicated by the heterogeneous of the sand and shale distribution and also by the lack of resolution when imaging thin channel deposits. Amplitude variation with offset (AVO) is a very important technique that is widely applied to locate hydrocarbons. Inaccurate estimates of seismic reflection amplitudes may result in misleading interpretations because of these problems in application to turbidite reservoirs. Therefore, an efficient, accurate, and robust method of modeling seismic responses for such complex reservoirs is crucial and necessary to reduce exploration risk. A fast and accurate approach generating synthetic seismograms for such reservoir models combines wavefront construction ray tracing with composite reflection coefficients in a hybrid modeling algorithm. The wavefront construction approach is a modern, fast implementation of ray tracing that I have extended to model quasishear wave propagation in anisotropic media. Composite reflection coefficients, which are computed using propagator matrix methods, provide the exact seismic reflection amplitude for a stratified reservoir model. This is a distinct improvement over conventional AVO analysis based on a model with only two homogeneous half spaces. I combine the two methods to compute synthetic seismograms for test models of turbidite reservoirs in the Ursa field, Gulf of Mexico, validating the new results against exact calculations using the discrete wavenumber method. The new method, however, can also be used to generate synthetic seismograms for the laterally heterogeneous, complex stratified reservoir models. The results show important frequency dependence that may be useful for exploration. Because turbidite channel systems often display complex vertical and lateral heterogeneity that is difficult to measure directly, stochastic modeling is often used to predict the range of possible seismic responses. Though binary models containing mixtures of sands and shales have been proposed in previous work, log measurements show that these are not good representations of real seismic properties. Therefore, I develop a new approach for generating stochastic turbidite models (STM) from a combination of geological interpretation and well log measurements that are more realistic. Calculations of the composite reflection coefficient and synthetic seismograms predict direct hydrocarbon indicators associated with such turbidite sequences. The STMs provide important insights to predict the seismic responses for the complexity of turbidite reservoirs. Results of AVO responses predict the presence of gas saturation in the sand beds. For example, as the source frequency increases, the uncertainty in AVO responses for brine and gas sands predict the possibility of false interpretation in AVO analysis.

Unit commitment is one of the most important subjects with respect to the economical operation of power systems, which attempts to minimize the total thermal generating cost while satisfying all the necessary restrictive conditions. ¡@¡@This thesis proposes a short-term thermal generating unit commitment by genetic algorithm and back propagation network. Genetic algorithm is based on the optimization theory developed from natural evolution principles, and in the optimization process, seeks a set of solutions simultaneously rather than any single one by adopting stochastic movement rule from one solution to another, which prevents restriction to fractional minimal values. Neural networks method outperforms in speed and stability. This thesis uses back propagation network method to complete neural networks and sets the optimal unit combination derived from genetic algorithm as the target output. ¡@¡@Under fixed electrical systems, instant responsiveness can be calculated by neural networks. When the systematical architecture changes, genetic algorithm can be applied to re-evaluation of the optimal unit commitment, hoping to improve the pitfalls of traditional methods. ¡@¡@This thesis takes the power system of six units for example to conduct performance assessment. The results show that genetic algorithm provides solutions closer to the overall optimal solution than traditional methods in optimizing unit commitment. On the other hand, neural networks method can not only approximate the solution obtained by genetic algorithm but also process faster than any other methods.

none

Abstract The traditional fault inspection of the motorcar engine cannot detect the noise and sound signal resulted from the abnormalities of some mechanical parts. For instance, the cylinder misfires; the looseness of the fan belt is irregular; the valve clearance is out of order¡K. and so on. When the fault message cannot be delivered by the ECU of the computer, the skilled senior engineers are required at this moment to make the experiential judgments. In the present society, due to the development of information, the computer technology makes progress by leaps and bounds. If we can make use of the monitoring method by the Acoustic signal instrument, build up a set of complete and efficient fault diagnosis system through the computer software and apply speedy and accurate way to assist the repairmen in relocating the causes for such faults, the accuracy of inspection can be greatly enhanced with a huge help in the preventive maintenance work. In that case, the fault conditions of the engine can be validated precisely and effectively, so the overhaul efficiency of the engine can be upgraded to a large extent. In this article, the procedures of sound signal recording will be brought forward by linking the digital camera with such a recording equipment as the high-precision microphone to make records of the fault sounds made when the engine runs. It uses the frequency analyzer to conduct the sampling and combine the computer software to further process and analyze the same. Finally the character parameters will be obtained. By applying the mathematical exercise of ¡§Back-Propagation Neural Network¡¨ to undertake the training and detection of the sounds for the purpose of identifying the kinds of the faults. It replaces the errors caused from the experiential judgments made by the expert senior engineers. In terms of the training and maintenance ability of the newly recruited technical repairmen, their capability for exact and reasonable recognition of the fault types is substantially promoted. Keywords¡GAcoustic Signal¡ABack Propagation Neural Network

No description available.

Thesis (M. Sc.)--University of Hong Kong, 1972. / Mimeographed.

Thesis (M.S.)--Brigham Young University. Dept. of Electrical and Computer Engineering, 2003. / Includes bibliographical references (p. 65-66).