Wilson H. Tang

Bayesian Framework for Characterizing Geotechnical Model Uncertainty

As any model is only an abstraction of the real world, model uncertainty always exists. The magnitude of model uncertainty is important for geotechnical decision making. If model uncertainty is not considered, the geotechnical predictions and hence the decisions based on the geotechnical predictions might be biased. In this study, a framework for characterizing geotechnical model uncertainty using observation data is proposed. The framework is based on the concept of multivariable Bayesian updating, in which the statistics of model uncertainty are updated using observed performance data. Uncertainties in both input parameters and observed data can be considered in the proposed framework. To bypass complex computational works involved in the proposed framework, a practical approximate solution is presented. The proposed framework is illustrated by characterizing the model uncertainty of four limit equilibrium methods for slope stability analysis using quality centrifuge test data. Parametric study in the illustrative example shows that both quality and quantity of the performance data could affect the determination of the model uncertainty, and that such effects can be systematically quantified with the proposed method.

Efficient Probabilistic Back-Analysis of Slope Stability Model Parameters

Back-analysis of slope failure is often performed to improve ones knowledge on parameters of a slope stability analysis model. In a failed slope, the slip surface may pass through several layers of soil. Therefore, several sets of model parameters need to be back-analyzed. To back-analyze multiple sets of slope stability parameters simultaneously under uncertainty, the back-analysis can be implemented in a probabilistic way, in which uncertain parameters are modeled as random variables, and their distributions are improved based on the observed slope failure information. In this paper, two methods are presented for probabilistic back-analysis of slope failure. For a general slope stability model, its uncertain parameters can be back-analyzed with an optimization procedure that can be implemented in a spreadsheet. When the slope stability model is approximately linear, its parameters can be back-analyzed with sensitivity analysis instead. A feature of these two methods is that they are easy to apply. Two case studies are used to illustrate the proposed methods. The case studies show that the degrees of improvement achieved by the back-analysis are different for different parameters, and that the parameter contributing most to the uncertainty in factor of safety is updated most.

Comparison of blue and green InGaN∕GaN multiple-quantum-well light-emitting diodes grown by metalorganic vapor phase epitaxy

InGaN∕GaN multiple-quantum-well (MQW) blue and green-light-emitting diodes (LEDs) were grown on sapphire substrates using metalorganic vapor phase epitaxy. High-resolution transmission microscopy shows that a much larger density of stacking faults exist in the quantum-well region of the blue LEDs than in the green LEDs. In the green LEDs, the blueshift in the electroluminescence (EL) emission energy at larger driving currents is more prominent than in the blue LEDs, which is explained by different strength of quantum-confined Stark effect as a result of different piezoelectric field intensity by different scales of strain relaxation in the blue and green MQWs. The steady broadening of the EL emission energy linewidth on the higher energy side with the increase of the driving current was observed in both blue and green LEDs, which is attributed to the band filling effect.

Profitability and welfare gain of private toll roads in a network with heterogeneous users

The recent worldwide tendency towards the introduction of commercially and privately provided roads proves to be efficient for building and operating road systems. One of the important issues concerning development of a private toll road is the selection of its capacity and toll charge and the evaluation of the relevant benefits to the private investor, the road users and the whole society. Yang and Meng [Transportation Research 34A (2000)] have recently examined this critical issue in a road network context. An important limitation in their analysis is that a single (average) value of time (VOT) for all road users is assumed in the network equilibrium model for traffic forecasts. It is a fact that each user has a different VOT, depending on how much money or time he or she is willing or able to spend on a particular trip. This paper relaxes the earlier single VOT assumption using a multi-class network equilibrium model. The entire population of users is segmented into a number of groups or classes according to their VOT, and trip rates by different groups are characterized by distinct group-specific demand functions. Users are assumed to minimize their individual generalized trip cost, and thus divide themselves among the various alternative routes that are differentiated on the basis of travel time and monetary cost. With the explicit consideration of user heterogeneity, we investigate the various possibilities of profitability and welfare gain of a private toll road in a given network under various combinations of road capacity and toll charge. In particular, we compare and contrast the outcomes with the case of a single average VOT, and investigate how VOT distribution affects traffic flow and profit forecasts of private toll roads.

Weathering indices for rhyolitic tuff and granite in Hong Kong

A series of chemical analyses and dry density tests of rhyolitic tuff and granitic samples have been carried out. The test results have been interpreted using a newly developed theoretical weathering model for calculating a mobility index of major elements in rock and soil. It is found that this index is useful for describing changes of rock and soil during weathering and classifying the degrees of weathering. Based on the theoretical model, a new parameter called the volume index, which combines chemical and physical data, is suggested. This index appears to be a good indicator for classifying the degree of weathering in Hong Kong. A quantitative classification scheme is suggested.

A Multiperiod Dynamic Model of Taxi Services with Endogenous Service Intensity

This paper presents a spatially aggregated multiperiod taxi service model with endogenous service intensity. The whole day service period is divided into a number of subperiods; during each subperiod, taxi supply and customer demand characteristics are assumed to be uniform. Customer demand is period-specific and described as a function of waiting time and taxi fare. Taxi operating cost for each work shift consists of two components: one component a function of total service time and the other component period-dependent. Each taxi driver can work for one or more shifts each day and freely chooses the starting and ending time of each shift. Equilibrium of taxi services is obtained when taxi drivers cannot increase their individual profits by changing their individual working schedules. A novel clock network representation is proposed to characterize the multiperiod taxi service equilibrium problem. The problem of interest is formulated as a network equilibrium model with path-specific costs and arc-capacity constraints, which can be solved using conventional nonlinear network flow optimization methods. The proposed model can ascertain at equilibrium the service intensity and utilization rate of taxis and the level of service quality throughout the day. The information obtained is useful for the prediction of the effects of alternative government regulations on the equilibrium of demand and supply in the urban taxi industry.

Planar integration of E/D-mode AlGaN/GaN HEMTs using fluoride-based plasma treatment

A planar-fabrication technology for integrating enhancement/depletion (E/D)-mode AlGaN/GaN high-electron mobility transistors (HEMTs) has been developed. The technology relies heavily on CF4 plasma treatment, which is used in two separate steps to achieve two objectives: 1) active device isolation and 2) threshold-voltage control for the enhancement-mode HEMT formation. Using the planar process, depletion- and enhancement-mode AlGaN/GaN HEMTs are integrated on the same chip, and a direct-coupled FET logic inverter is demonstrated. Compared with the devices fabricated by a standard mesa-etching technique, the HEMTs by a planar process have comparable dc and RF characteristics with no obvious difference in the device isolation. The device isolation by a plasma treatment remains the same after 400 degC annealing, indicating a good thermal stability. At a supply voltage (VDD) of 3.3 V, the E/D-mode inverters show an output swing of 2.85 V, with the logic-low and logic-high noise margins at 0.34 and 1.47 V, respectively

Reliability-Based Optimization of Geotechnical Systems

Reliability-based optimization (RBO) intends to minimize the cost involved in a design while satisfying all technical requirements specified by using reliability theory. Although this idea is attractive, its implementation is generally difficult because an accurate evaluation of the reliability constraints is often computationally prohibitive when it is coupled with the minimization of a cost function. This paper suggests an indirect method for RBO of geotechnical systems initially on the basis of the mean first-order reliability method (MFORM). Because reliability analysis with MFORM is very efficient, the computational work involved in RBO is greatly reduced. The major concern about MFORM is that it is not an accurate reliability method. A reliability index mapping function was used to relate the reliability index calculated by MFORM to that calculated by a more accurate method. When the mapping function perfectly reflects the relationship between MFORM reliability index and that from an accurate reliab...

Slope Reliability Analysis Considering Site-Specific Performance Information

The performance of a slope, such as surviving a certain groundwater condition, can be viewed as the outcome of a full-scale test performed directly on the slope and may provide valuable information for safety assessment, upgrading analysis, and repair design of the slope. Performance information can be divided into two types: (1) the slope survived a certain state, and (2) the slope failed at a certain state. This paper illustrates two methods for slope reliability analysis considering site-specific performance information, i.e., an indirect method based on back-analysis of the performance information, and a direct method, in which the back-analysis procedure is bypassed. The two methods are theoretically the same but different in implementation details. As examples, an existing slope for safety assessment and a failed slope to be repaired are studied in this paper. Considering the past survival information increases the reliability of the slope. The increase in reliability is larger if the slope survives a more critical state. Thus, ignoring the survival information may result in uneconomical decisions. In contrast, ignoring the failure information may either underestimate or overestimate the reliability of a slope. As a result, neglecting the past failure information may lead to unsafe or uneconomical decisions.

Uncertainties of Field Pullout Resistance of Soil Nails

A large number of field pullout tests on soil nails have been carried out to provide valuable information for enhancing the understanding of pullout resistance of soil-grout interface and for reliability evaluation of soil-nailed slopes. In this paper, a data set of 167 field pullout tests performed in 23 nailed completely decomposed granite cut slopes is used for a statistical evaluation of four factors influencing the pullout resistance of soil nails, namely overburden pressure, grout length, soil suction, and soil dilatancy. For the tests in which nails were pulled out, the measured pullout resistance is essentially independent of the effective overburden pressure. A bias factor r* is defined as the ratio of the measured pullout resistance and the calculated value using a design equation. The mean value of r* is 4.30 and the coefficient of variation is 47%. When the uncertainties in grout length, soil suctions around nails, and the soil shear dilatancy are considered, the mean value of r* can finally b...