1、附件 Analytic Element Method The research in groundwater flow isaimed at producing tools capable of accurate and efficient modeling of large regionalsystems of groundwater flow. These systemsoften consist of many water-bearing layers,called aquifers, that are separated by leakylayers that resist flow,
2、 but are not impermeable. Most groundwater enters such aquifersystems via the uppermost layer in the formof infiltration from rainfall. Accurateprediction of contaminant movement andanalysis of the zone to be protected forwells (the well-head protection plan)requires accurate modeling of the leakage
3、between the aquifers and through the leakylayers. The method developed at this department for such regional groundwater modeling is the Analytic Element Method. The twomost important applications of this method, to problems of the kind described above,are the Metropolitan Groundwater Model (Metro Mo
4、del) of the Twin CitiesGroundwater Basin (seehttp:/www.pca.state.mn.us/water/groundwater/metromodel.html), and the Dutch National Groundwater Model NAGROM. The Analytic Element Method is based upon the superposition of analytic functions,called analytic elements, which can be adjusted to meet certai
5、n boundary conditions. Themethod can be applied to general problems of two-dimensional and three-dimensionalflow. Current research focuses on the accurate modeling of leakage between aquifers, the combination of two-dimensional and three-dimensional elements in large regionalmodels, and the so-calle
6、d superblock approach which makes it possible to create verylarge computer models that can produce results quickly. The two-dimensional Analytic Elements are developed using an approach known asWirtinger Calculus, which is based on a change of variables that permits all analysis to be carried out in
7、 the complex domain, regardless of the nature of the vector field thatdescribes the flow. The vector field in groundwater flow represents the discharge in eachaquifer； the properties of this vector field, i.e., its divergence and curl, depend on thenature of the flow. The Analytic Elements are selec
8、ted such that some create a vectorfield that is divergence-free and irrotational (Laplace Field), others create divergence, andstill others rotation. Figure 1: Contours of constant displacement components for a uniformly stressed elastic medium with inclusions of different elastic properties. These
9、calculations were performed using the Analytic Element Method, initially developed for groundwater flow. Figure 2: Capture zone formed by streamlines emanating from a well for a waste repository. The area containing waste is surrounded bya slurry wall (a trench filled with a material with high resis
10、tance to flow). A small well inside the area captures whatever rainfall enters thearea and may be polluted. The figure on the left corresponds to the case for which the wall is designed, whereas the figure on the right showsthe case when a small opening exists, causing significant leakage to occur.
11、It is of interest to note that for both cases the well draws atmaximum capacity, with the head in the well screen set to the base of the aquifer. The discharge for the case of the faulty wall is about 200times as large as for the case of the intact wall. Acoustic Emission Monitoring A common feature
12、 of failure for rock is the development of microcracking, whichreleases energy in the form of elastic waves called acoustic emission (AE). The AE technique can be used to monitor the evolution of damage, through the entire volume, atvarious stages of loading. The locations of AE provide a picture th
13、at forms a basis forthe justification of mechanical models of damage and failure. Indeed, a physical interpretation of the data may be used to define a characteristic length of a quasi-brittle material. In modeling the response of rock, the characteristics of the localized damage zonemay be importan
14、t in predicting failure. Figure 3: Locations of acoustic emis- sion up to about 95% of the peak stress (blue) and around peak stress (red) in a beam test with a Berea Sandstone. Microcrackingwas more or less random prior to peak stress, but a localized region can clearly be identified at peak stress
15、. Nondestructive Pavement Characterization With our highway systems deteriorating, their timely monitoring and repairs areessential. Currently, a number of seismic testing devices such as the falling weightdeflectometer are available for non-intrusive pavement diagnosis. The effectiveness ofsuch rem
16、ote sensing tools remains intrinsically hampered, however, by the difficulty ofdata interpretation. In this investigation, an advanced back-analysis is developed fordelineating the pavement subgrade profile from FWD measurements. The inverse solution is based on an artificial neural network approach
17、 as a pattern recognition tool, and a viscoelastodynamic pavement model as a predictive device. To expose thepavementsresonance and stiffness characteristics, ground deflection data are interpreted in terms ofsuitable frequency response functions. Robustness of the back-analysis is improved via(i) a
18、n integral data sampling scheme, (ii) noise injection technique used in the neuralnetwork development, and (iii) low-pass filtering of the seismic records, employed tominimize the masking effect of lateral wave reflections on the pavement edges. Figure 4: Comparison of the theoretical frequency resp
19、onse function, produced by the back- analysis, with the field measurements taken at the Minnesota Road Testing facility. Longitudinal Cracking of Flexible Pavements The cracking along the wheel path (longitudinal) is of major concern in flexible pavements. These cracks originate at the surface of the asphalt layer and often terminatebefore reaching the base of the pavement. To gain an insight into the potential mechanisms of the formation of longitudinal cracking, numerical simulations are carried
