Best Selling Books by Lin Wang

Application-level fault tolerance incurs significant development-time cost. FATMAD is a fault-tolerant multi-agent development framework that is built on top of a mobile agent platform (Jade) and provides development support for application level fault tolerance. FATMAD aims to satisfy the needs of two communities of users: agent application developers and fault-tolerant protocol developers. FATMAD is based on a generic fault-tolerant protocol whose refinements lead to a broad range of checkpoint and recovery protocols to be used in user applications. This, in turn, can significantly reduce the development time of fault-tolerant agent applications. FATMAD allows application developers to apply suitable protocols and customizable deployment configurations to their applications so that the characteristics and fault tolerance requirements of a particular application can be addressed. FATMAD implements common facilities that are required by different protocols. The protocol-specific parts in each protocol can be extended by protocol designers.

Asphalt with high viscosity, air, and water constitutes multi-phase turbulent flow accompanied by complex dynamic phenomena such as heat transfer and phase transition in the foaming chamber during the asphalt foaming process. The foaming process is difficult to fully describe by mathematical model. In order to establish the correspondence between the approximate simulation flow field in the foaming chamber and the physical experiment, it is necessary to propose a reasonable evaluation method for the flow field simulation in the foaming chamber. Approximate numerical expression of asphalt foaming process is established by the volume of fluid (VOF) multiphase flow calculation method combined with self-defined phase and mass transfer functions. Based on the principle of asphalt foaming, effective foaming area is defined as the area where temperature is =100°C, volume fraction of asphalt is =0.5, and volume fraction of water vapor is between 0.1 and 0.4. The ratio of the effective foaming area to the total flow field is taken as the evaluation index. The proportion of effective foaming area obtained by asphalt foaming numerical simulation is compared with the expansion rate and half-life obtained by asphalt foaming experiments with three foaming chambers of different sizes. The comparison shows that ratio of effective foaming area has the same tendency as expansion rate and half-life. And the simulation evaluation index is verified. Suggestions are put forward on the structure design of asphalt foaming chamber based on the distribution of effective foaming area. It shows that this method achieves rapid theoretical evaluation of foaming chamber structure and could be used for the optimal design of foaming chamber.

The behavior of cement mortar incorporating fly ash varies with chemical and physical characteristics of fly ash. Of many parameters affecting the compressive strength of fly ash mortar, five were investigated in this study. They are the particle size, the particle size distribution, fineness of fly ash, the CaO content and the Fe2O3 content. Fly ashes from two different sources were used in this study. One comes from the Public Service Electric and Gas (PSE & G)''s electrical generation station in Hudson County and the other is from the PSE & G plant in Mercer County. Each was separated into seven different particle size ranges. The results of the tests indicated that there are direct relationships between the particle size, Blaine fineness, mean diameter of fly ash and the compressive strength. Fe2O3 content of fly ash is not found to have a notable effect on the mortar strength. CaO content of fly ash varying from 2.47% to 6.76% also has no significant effect on the strength of mortar. Additionally, the effect of chromium on cement mortar and concrete was studied. Trivalent and haxavalent [sic] chromium were used in the experiment. The compressive strength of mortar and concrete incorporating Cr(III), as well as mortar and concrete incorporating Cr(VI) were tested up to 180 days according to ASTM C-109. Leaching tests were conducted with different pH extractants. The results of the tests showed that the leaching characteristic of Cr(III) mortar and Cr(VI) mortar are different. The influence of Cr(III), and Cr(VI) on compressive strength of mortar and concrete is also varied. The results of leaching tests indicated that cement is very good for immobilizing Cr(III) under field condition, unless the pH is extremely low. However, Cr(VI) can be leached from the mortar at early ages. The compressive strength of Cr(III) mortar and concrete is higher than that of the conventional mortar and concrete at all ages. But the strength of Cr(VI) mortar and concrete is lower than that of the conventional mortar and concrete. A major finding was the discovery of significant amount of soluble chromium in Mercer fly ashes. The results show that the majority of the chromium in Mercer fly ash is concentrated in the small particles, those in the 0-10 micron range.

Rural childhood obesity is a neglected but pressing issue in the United States. The childhood obesity prevalence in rural areas is significantly higher than that in urban areas, and the higher risk linked to the rural context is not sufficiently understood. This dissertation explores and dissects the relationship between rurality and childhood obesity risk by analyzing nationally representative data. Three main datasets are used in this research. These include two county-level datasets: the American Community Survey (ACS) data provides sociodemographic characteristics and the U.S. Census Bureau County Business Patterns data offers food environment variables. The Early Childhood Longitudinal Study-Kindergarten Cohort 2011 (ECLS-K: 2011) provides individual-level data. These data sources are merged and analyzed to assess hypothetical associations suggested by a multilevel conceptual framework. The analysis has two components. First, the study employs multilevel modeling to examine correlates of body mass index (BMI) among three age cohorts of children in the cross-section. Second, the study presents estimates of growth curve models to estimate the baseline and growth of BMI as a function of time-variant and time-invariant variables. The results indicate that although childhood obesity prevalence is higher in the rural US., rurality does not appear to have a direct effect on childhood obesity risk. Contextual factors such as poverty rate at the county level and food environment have significant associations with BMI status and change that help account for the positive association of rural residence. At the individual level, parental education and childrens daily activities have direct impacts on childhood BMI and obesity risk. At the environmental level, access to less healthy food retailers emerges as a key factor influencing BMI. There is no interaction effect between individual variables and rural residence. Food environment is also found to affect BMI growth, and the effect size increases over time. In sum, the findings suggest that rural residence plays a minor role in childhood obesity risk. On the other hand, the lower socioeconomic status and disadvantaged food environment more common to rural areas places rural children at a higher risk for obesity. Several policy responses are suggested, including providing nutrition education that emphasizes the causes and consequences of obesity, and increasing access to a healthy and affordable food supply to counteract the implications of economic deprivation and poor food environments.

Fluid-borne noise is one of the main components of hydraulic noise. Its attenuation may have a significant effect on the cost of hydraulic systems. Standard passive silencers and dampers can be useful in reducing it in certain frequency ranges; however, these tend to be heavy, bulky and expensive. Active control algorithms, which are a comparatively recent means of reducing fluid-borne noise, can be applied to overcome this compromise. The work presented in this thesis is the development of some active control algorithms utilized in a simple hydraulic system to cancel a number of harmonic orders of fluid-borne noise generated by a servo valve or a real pump. To realize cancellation the filtered reference least mean square (FXLMS) adaptive control method is mainly presented. Furthermore, a fast response servo valve is applied as an actuator to generate a proper anti-noise flow signal in real-time. For simplicity, an off-line identification method for the secondary path is applied in the time invariant working condition. Moreover, ripple reflection from both ends of the hydraulic circuit can produce different effects under different working conditions. In order to execute the cancellation without any prior information about the dynamics of hydraulic systems, the on-line secondary path identification method is discussed. However, in this algorithm an auxiliary white-noise signal applied to an on-line method may contribute to residual noise and an extra computation burden may be added to the whole control system. The performance of these control algorithms is firstly investigated via simulation in a hydraulic pipe model and the real-time application on a test rig using a servo valve as a noise source. Finally, these schemes are realized in a simple hydraulic system with a real pump noise source. The fluid-borne noise can be attenuated by about 20 dB in normal working conditions.

Abstract - if sent.