@phdthesis{Zhang2015,
  author    = {Guowei Zhang},
  title     = {A well generalization method for multiple production wells with variable discharge applied to a coal pit in An-Hui province of China},
  journal    = {Eine verallgemeinerte Methode zur Auswertung von Pumptests an multiplen F{\"o}rderbrunnen mit variabler Abflussmenge am Beispiel einer untert{\"a}gigen Kohlemine in der Provinz An-Hui, China},
  url       = {https://nbn-resolving.org/urn:nbn:de:gbv:9-002261-1},
  year      = {2015},
  abstract  = {Liu\–II coal pit is a typical example of China\’s deep coal mines which is seriously threatened by groundwater inrush from the underlying carboniferous Taiyuan limestone formation. An exhaustive data set of this confined aquifer exists. The aquifer lies 45 m\∼ 60 m below the major coal seam. A traditional artesian aquifer test has been performed in order to assess the hydraulic properties, e.g. transmissivity (T) and storage coefficient (S). This artesian aquifer test is conducted with four simultaneously operating production wells while the discharge of each production well varied with time. The results of this test suggest that the aquifer is heterogeneous. Therefore, the according problems are: (1) how to analyze the artesian aquifer test with linearly declining discharge; (2) how to deal with multiple production wells in an aquifer test; (3) how to adequately consider aquifer heterogeneity. Thus, the objective of this thesis is to solve these problems. 1) As opposed to classical above-ground pumping tests, it is difficult to control the discharge rate of the production well in a deep mine artesian aquifer test since the hydraulic pressure is extraordinary high. Moreover the discharge rate won\’t descend rapidly to zero, thus the analytical solution of Jacob and Lohman (1952) type curve for the artesian aquifer test will not be applicable. It is more reasonable to analyze the test as a pumping test with variable discharge. It is considered to rebuild a hydrogeological conceptual model which is similar with Theis (1935) model but with the variable discharge. A general equation for any discharge variability is given. Its application for the linearly declining discharge is presented subsequently, and a type curve of this equation with linearly declining discharge is given as well. After that, a simple numerical model is built by FEFLOW to simulate an artificial pumping test with the linearly declining discharge by assigning different parameter sets for transmissivity and storage coefficient. The type curve method is applied to evaluate transmissivity and storage coefficient for the linearly declining discharge well. The deviation between the given values of transmissivity and storage coefficient in FEFLOW and the values of those calculated by matching point are sufficiently small. Thus, when the discharge of production well declines linearly, a type curve method as an empirical method is reasonable and gives satisfactory values of these hydrogeological parameters. 2) In some cases, it is necessary to conduct a pumping test (or an artesian aquifer test) with several pumping wells (or production wells) which work simultaneously in order to discharge maximum quantity of groundwater. Normally, the superposition method or numerical simulation is applied to analyze the test result. However, a new approach called \“Well Generalization Method\” is defined and analyzed in this thesis. It is an easy\–to\–use approach for hydrogeologist to estimate the aquifer parameters while conducting an aquifer test. Since the key point of this approach is using a generalization well to substitute the pumping (or production) wells, it is obvious that this approach will generate the estimated error of parameters. Accordingly, several scenarios are analyzed and discussed based on the artificial type aquifer designed in FEFLOW. A homogeneous aquifer and a heterogeneous aquifer which is generated by geostatistical stochastic simulation technique (see 3)) are discussed separately. As a result, this approach is feasible and applicable under some conditions when the calculated observation well is arranged more than about 2.5 times the scale of the multi\–pumping\–wells field away from the center of the multi\–pumping\–wells field, furthermore, the maximum deviation of drawdown resulting from these observation wells will be less than 0.5 m, and the estimated value of transmissivity will be 0.44\% smaller than real value. 3) Finally aquifer heterogeneity is addressed, in order to check the introduced method for applicability under realistic conditions. It has been described that aquifer heterogeneity plays a major role in hydrodynamic processes (e.g. de Marsily et al., 1998). Geostatistics which is considered as a useful tool for characterizing the spatial variability of transmissivity is applied to solve this problem. Based on the results of the artesian aquifer test conducted in Liu\–II coal pit, a model of spatial variability of transmissivity is developed. Sequentially, the variogram model is applied in ordinary kriging to interpolate the transmissivity distribution, and in sequential Gaussian simulation to simulate a random field of transmissivity data in order to reflect its small scale variability. A comparison of the results of estimation and simulation of transmissivity indicates that the simulated values better reflect the spatial variability, reversely, the estimated values are much smoother.},
  language  = {en}
}