汶上县地下水流数值模型及其替代模型研究

发布时间：2018-11-05 14:55

【摘要】：地下水管理模型是用来解决在某些约束前提下,通过对模型决策变量的操控,使得地下水系统根据规定目标获得最优化的一种手段。在求解这种优化模型时,需要反复调用数值模拟模型,进而导致求解地下水管理模型的时间大大增加,制约了地下水管理模型的发展与实际应用。替代模型可以在实现数值模拟模型功能的前提下,大大简化计算,从而推动地下水管理模型在实际工作中的应用。汶上县地处鲁西平原和鲁中低山丘陵的交接地带,是著名的国家商品粮基地县,盛产多种经济作物。在汶上县的用水结构中,地下水资源占了较大的比例,尤其是城乡生活饮用水及农业灌溉用水。因此,地下水资源对于汶上县的发展起着举足轻重的作用。随着汶上县人口规模和经济体量的增长,所需要的水资源量也急剧增加,进而导致地下水资源遭到过度开采,造成了地下水位持续下降等地质环境及社会经济问题。因此,地下水资源的供需矛盾制约着汶上县的经济发展,进而也对今后的水资源管理工作提出了更高要求。本文通过收集汶上县的地质以及水文地质方面的基础资料,建立了水文地质概念模型,并以此为基础使用GMS模拟软件建立了汶上县地下水流数值模型。通过拟合实际流场以及长观孔,使得所构建的数值模型能够更逼近真实水文地质条件。使用蒙特卡罗抽样方法及拉丁超立方抽样方法在汶上县中部十个乡镇地下水利用量的适宜范围内,分别抽样确定了200组及20组地下水开采方案。同时,以地下水流现状模型为基础,建立2015年到2030年的预测模型,计算在各组地下水开采方案下,研究区内地下水水位降深的平均值。根据蒙特卡罗方法和拉丁超立方抽样方法确定的地下水开采方案及其对应的地下水水位降深分别作为训练样本和检验样本,使用BP神经网络及RBF神经网络构建了汶上县地下水流数值模型的替代模型,并对其有效性进行检验。结果表明,使用所建立的BP神经网络模型及RBF神经网络模型得出的地下水水位降深与根据预测模型得出的降深基本接近。因此,两种神经网络模型可以作为汶上县地下水流数值模型的替代模型来使用,从而避免了在地下水管理模型中反复运行数值模型,提高了管理模型的运行效率。
[Abstract]:Groundwater management model is used to solve the problem that groundwater system can be optimized according to the specified objectives by manipulating the decision variables of the model under some constraints. In order to solve this kind of optimization model, the numerical simulation model needs to be called repeatedly, which leads to a great increase in the time of solving groundwater management model, which restricts the development and practical application of groundwater management model. The substitution model can greatly simplify the calculation on the premise of realizing the function of numerical simulation model, thus promoting the application of groundwater management model in practical work. Wenshang County is a famous commodity grain base county, which is located in the transition zone between Luxi Plain and low mountain hills in the middle of Shandong Province. It is rich in many kinds of cash crops. In the water structure of Wenshang County, groundwater resources account for a large proportion, especially urban and rural drinking water and agricultural irrigation water. Therefore, groundwater resources play an important role in the development of Wenshang County. With the increase of population scale and economic volume in Wenshang County, the amount of water resources needed increases rapidly, which leads to the overexploitation of groundwater resources, resulting in geological environment and social and economic problems such as the continuous decline of groundwater level. Therefore, the contradiction between supply and demand of groundwater resources restricts the economic development of Wenshang County, and puts forward higher requirements for water resources management in the future. Based on the basic data of geology and hydrogeology in Wenshang County, a conceptual model of hydrogeology is established in this paper, and a numerical model of groundwater flow in Wenshang County is established by using GMS software. By fitting the actual flow field and the long hole, the numerical model can approach the real hydrogeological conditions more closely. Using Monte Carlo sampling method and Latin hypercube sampling method, 200 groups and 20 groups of groundwater exploitation schemes were determined by sampling within the suitable range of groundwater utilization in ten towns in the central part of Wenshang County. At the same time, based on the groundwater flow status model, a prediction model from 2015 to 2030 is established to calculate the average groundwater level drop in the study area under each groundwater exploitation plan. According to the groundwater exploitation scheme determined by Monte Carlo method and Latin hypercube sampling method, and the corresponding groundwater level drop, they are used as training samples and test samples, respectively. The BP neural network and RBF neural network are used to construct a numerical model of groundwater flow in Wenshang County and its validity is tested. The results show that the groundwater level depth obtained by using the established BP neural network model and the RBF neural network model is basically close to that obtained from the prediction model. Therefore, the two neural network models can be used as the replacement model of Wenshang groundwater flow numerical model, thus avoiding repeated operation of numerical model in groundwater management model and improving the efficiency of management model.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P641.8

【相似文献】