Chiu-Shia Fen

A web-based distributed problem-solving environment for engineering applications

In this study, a prototype of web-based distributed problem-solving environment (W-DPSE) is presented to facilitate computer aided engineering (CAE) technologies. This system provides an effective approach to distributed modeling and simulation, and in addition, to support networked collaboration such that scientists around the world could interactively, visually and experimentally explore their daily design work through the proposed system. The W-DPSE system is constructed as a three-tiered architecture, consisting of three major layers: a web client presentation interface (WCPI), computing solver servers (CSS), and a system management server (SMS). All the components within this architecture are implemented with an object-oriented approach--Java while transparent communication across these three layers is attained through the remote method invocation (RMI) technology. To generalize the applicability of the system so as to accommodate interdisciplinary and diversified applications, a novel and efficient interface is developed for wrapping legacy computation codes, including C and FORTRAN etc., as a Java component. Note that in the interface, the communication mechanisms between Java objects and legacy codes are implemented by way of java native interface (JNI) and Unix inter-process communication (IPC) provided by OS. At last, the applicability of the W-DPSE system is extensively confirmed through the practices of two engineering applications: topology optimization of structures and pollutant transport simulations of coastal waters.

Effect of Well Radius on Drawdown Solutions Obtained with Laplace Transform and Green’s Function

We have developed a drawdown solution for a partially penetrating well under constant flux pumping in a confined aquifer with finite thickness. The predictions of our solution diverge from the predictions of Hantush’s solution (1961), particularly for problems with low ratios of well screen length to aquifer thickness. Furthermore, the predicted drawdown from Hantush’s solution (1961) differs from that of Yang et al.’s solution Water Resour Res 42:W0552, (2006) only near the well and at small time values as indicated in Yang et al. Water Resour Res 42:W0552, (2006). Our solution is based on Green’s function with a columnar source (sink) that represents pumping from a finite-radius well. Hantush’s solution (1961) and Yang et al.’s solution Water Resour Res 42:W0552, (2006), however, were derived from Laplace transform techniques for pumping in a well with an infinitesimal and a finite radius, respectively.

Assessing the stormwater impact to an urban river ecosystem using estuarine water quality simulation model

The reduced richness of fish species in the Love River has long been associated with the impact of storm flooding. During the 3-year time period of 1993–1995, it has been observed that the devastating conditions of low DO levels in the Love River frequently appear at a certain time after the emergence of peak rainfall. While the official investigation with regard to estuarine and river ecology indicates that more than 85 fish species could be found in the neighbouring river systems in South Taiwan, only 17 fish species appeared in the estuarine system of the Love River. To integrate the knowledge of environmental and ecological impacts in these extreme storm events, a two-dimensional numerical model – Love River Hydrodynamic and Water Quality model (LRHWQ) – was developed to simulate the fate and transport of contaminants in the Love River estuarine system. It eventually produces the responsive information of water quality and the dissolved oxygen deficit with respect to the combined sewer overflow (CSO) impacts along the river corridor. To justify the impact of DO level on the fish community, a laboratory analysis was conducted with respect to the living threshold of young Liza macrolepis and Chanos chanos that are two representative fish species in the group of abundant species in the Love River. The LC50 found in this study lies in between 0.8–1.1 mg/L; yet the saturated oxygen concentration is normally near 9.0 mg/L in the study area. The fish community is sensitive to the dissolved oxygen deficit and simulation studies confirm that simply building a sewage interceptor system to handle routine wastewater effluents without regard to the CSO issues in the storm events cannot fulfil the ultimate goal of environmental restoration in this urban river system.

Modeling volatilization of residual VOCs in unsaturated zones: A moving boundary problem

It is of practical interest to investigate the natural evaporation of volatile organic compounds (VOCs) after the removal of a leaking tank situated on the top of the soil. This study aims to develop a mathematical model to predict mole fraction distributions and migration of evaporation front for two VOCs emanating from residual non-aqueous phase liquid (NAPL) due to the leak from the tank in a homogeneous soil. Considering the location of the front and the regions above and below the front, a numerical model for the diffusive transport of VOCs in unsaturated soils was developed using the finite difference method with a moving grid approach. The model was further simplified to the case of single VOC and solved analytically by Boltzmanns transformation with a moving boundary. Analytical expressions for the depth and moving speed of the front for a single VOC were then obtained for practical use. Finally, the developed model was used to predict the concentration distributions of VOCs below the land surface and examine the factors affecting the location and moving speed of the evaporation front.

An effective response surface-based optimisation approach for soil vapour extraction system design

While soil vapour extraction has been widely utilised as a remediation technology over the past decade, design and associated process performance modelling of full-scale systems has become important. This paper presents an effective optimisation approach, based on Zheng and Wang (Zheng, C. and Wang, P.P., 2002. A field demonstration of the simulation optimization approach for remediation system design. Ground Water, 40 (3), 258–265), for soil vapour extraction system (SVES) design. This approach integrates a multiphase flow and contaminant transport simulator, a genetic algorithm (GA) and a response surface (RS) method that employs non-linear regression analysis for constructing an approximation function as the objective of SVES constraint optimisation problems. The optimisation problem is to maximise the percentage of contaminant mass removal in terms of extraction rates at prescribed well locations under certain constraints. To demonstrate the effectiveness of the current approach, a few practical applications are performed and the associated optimal designs are compared against those using the simulation/optimisation approach. Results show that the proposed approach can provide a good optimal solution with a small computational effort.

Modeling wastewater discharge to the Ai estuary, Kaohsiung, Taiwan

Abstract A comprehensive modeling study was undertaken to examine the hydrodynamics and transport and fate of biochemical oxygen demand (BOD) and dissolved oxygen (DO) in the Ai Estuary, Kaohsiung, Taiwan. Especially, a laterally averaged hydrodynamic and water quality model was developed and calibrated in an attempt to understand the effluent impact on the water quality of the estuary. A set of hydraulic parameters based on two different field surveys was derived from the model calibration. However, some of the water quality parameters derived from different field data sets are not consistent. The parameter sensitivity analysis further shows that the BOD removal and deoxygenation rates and the chlorophyll a level, which all vary widely with time and space in the estuary, have significant effects on the BOD and/or DO dynamics. This reveals the necessity of investigating the range of the BOD rate constants and the seasonal variation of the Chl a level for the estuary. The model was then applied to wastewater impact assessment by evaluating the time interval for the DO level restored to 5 mg/L along the estuary under different wastewater discharge conditions. The simulation result is consistent with the field surveys. This study further reveals that the restoration time can be reduced considerably if the dischargerate and duration are moderately controlled.

Influence of spatial variability of soil properties on mass removal for a soil vapor extraction system

Abstract A multiphase flow and transport model is modified and applied to study the influence of spatial variability of soil properties on predictions of contaminant mass removal for a soil vapor extraction system (SVES). Spatial distributions of mean particle size (dm), soil porosity (ϕ), permeability (k) and interphase mass transfer rate between air and organic liquid phases (Kgo) for nonuniform soil composed of fine‐ to coarse‐grain sands are considered in the model. For these parameters, <p and dm are generated based on a geostatistical model and k and Kgo are determined according to theoretical or empirical correlations expressed in terms of dm and/or <p. Results show that uncertainty in spatial distribution of any of these parameters can exert a significant influence on the predicted removal percentage, even for cases with low variability in dm . In particular, to rigorously predict SVES contaminant removal efficiency, one must assess in‐situ Kgo value or the relationships of Kgo to other factors. The present simulations show that uncertainty in spatial variation of these parameters can result in a wide variety of organic liquid saturation distributions during SVES operation, even if the predicted removal percentages are similar for different cases. For long‐term soil remediation by SVE, in order to promote SVES mass removal efficiency, intermediate stages of the operation must include reassessment of in‐situ contaminant mass distribution. This new information can be used to relocate extraction wells, if necessary, to promote SVES mass removal efficiency.

Density-driven transport of gas phase chemicals in unsaturated soils

Variations of gas phase density are responsible for advective and diffusive transports of organic vapors in unsaturated soils. Laboratory experiments were conducted to explore dense gas transport (sulfur hexafluoride, SF6) from different source densities through a nitrogen gas-dry soil column. Gas pressures and SF6 densities at transient state were measured along the soil column for three transport configurations (horizontal, vertically upward and vertically downward transport). These measurements and others reported in the literature were compared with simulation results obtained from two models based on different diffusion approaches: the dusty gas model (DGM) equations and a Fickian-type molar fraction-based diffusion expression. The results show that the DGM and Fickian-based models predicted similar dense gas density profiles which matched the measured data well for horizontal transport of dense gas at low to high source densities, despite the pressure variations predicted in the soil column were opposite to the measurements. The pressure evolutions predicted by both models were in trend similar to the measured ones for vertical transport of dense gas. However, differences between the dense gas densities predicted by the DGM and Fickian-based models were discernible for vertically upward transport of dense gas even at low source densities, as the DGM-based predictions matched the measured data better than the Fickian results did. For vertically downward transport, the dense gas densities predicted by both models were not greatly different from our experimental measurements, but substantially greater than the observations obtained from the literature, especially at high source densities. Further research will be necessary for exploring factors affecting downward transport of dense gas in soil columns. Use of the measured data to compute flux components of SF6 showed that the magnitudes of diffusive flux component based on the Fickian-type diffusion expressions in terms of molar concentration, molar fraction and mass density fraction gradient were almost the same. However, they were greater than the result computed with the mass fraction gradient for >24% and the DGM-based result for more than one time. As a consequence, the DGM-based total flux of SF6 was in magnitude greatly less than the Fickian result not only for horizontal transport (diffusion-dominating) but also for vertical transport (advection and diffusion) of dense gas. Particularly, the Fickian-based total flux was more than two times in magnitude as much as the DGM result for vertically upward transport of dense gas.

Solution for Soil Vapor Extraction from a Pressure-Controlled Well

AbstractThis paper presents a new solution for a soil venting problem, in which a vertical, partially penetrating well is operated at a fixed pressure under covered, uncovered, or leaking ground su...

Estimating groundwater velocity using apparent resistivity tomography: A sandbox experiment

The electrical resistivity tomography (ERT) technique can estimate groundwater velocity to within 5% of the pre-set groundwater velocity. The apparent conductivity obtained by the ERT technique is linearly related to the groundwater conductivity, as described by Archies law. Gaussian-like profiles of the tracer concentration were demonstrated with the ERT technique, and the estimated dispersion coefficient was between 0.0015 and 0.0051 cm2/sec. In terms of monitoring changes in groundwater conductivity, the ERT technique has two major advantages over monitoring wells: (1) it measures a larger area and provides more representative results; and, (2) it does not withdraw groundwater samples, and therefore does not affect the groundwater flow. The objective of this research is to measure groundwater velocity with the ERT technique using only one well. The experiments in this research were divided into two parts. The first part evaluated the accuracy and repeatability of the ERT technique using a dipole-dipole array, and the second part estimated the groundwater velocity in a sandbox using the ERT technique. The length, width, and height of the sandbox, which was made of acrylic, were 1.5, 5, and 1.0 m, respectively. The ERT sandbox was sequentially filled with 5-cm layers of the silica sand to a total height of 70 cm. A total of 32 electrodes spaced every 5-cm were installed in the center of the sandbox. Three monitoring wells were installed along the line of the electrodes. Both no-flow and constant flow (NaCl solution with electrical conductivity and concentration of 5,000 μs/cm and 2.456 g/L, respectively) tracer experiments were conducted.