共查询到19条相似文献,搜索用时 156 毫秒
1.
对海底金矿床开采过程中不同高度岩层位移进行了监测,对岩层变形时间序列重构相空间,用混沌理论揭示了不同高度岩层位移在相空间中的相点距离演变规律。用神经网络建立了岩层变形相空间相点距离演化预测模型,预测了新立矿区海底开采岩层变形,并建立了海底开采岩层变形安全预警系统。采用梯度下降法与混沌优化方法相结合方法训练神经网络,使神经网络预测模型实现快速训练的同时,避免陷入局部极小,同时提高了模型计算精度。研究表明,岩层变形表现出混沌特征,对其相空间重构后,岩层变形的细微变化特征被放大,其内在规律能得到充分展示,为建立海下开采安全预警系统提供了基础。 相似文献
2.
边坡的位移预测对其稳定性的预报具有十分重要的意义,从基于相空间重构的BP神经网络预测方法对位移时间序列进行了分析,对相空间重构的参数延迟量以及嵌入维数进行了论述,将预测结果与传统的BP神经网络模型的预测结果进行了比较。结果表明,基于相空间重构的BP神经网络具有更高的精度,是一种优秀的预测方法。 相似文献
3.
4.
5.
根据1964.01—1978.12新乡百泉泉水流量动态分析,泉水流量动态受大气降水控制,反映了天然状态下的泉水流量动态特征,并建立了百泉逐月泉水流量的逐步回归预测模型。用BP神经网络方法,建立了不同时段的逐月泉水流量动态预测模型,并与逐步回归模型进行比较,BP神经网络模型拟合效果优于逐步回归模型。根据逐步回归模型和BP神经网络模型预测的结果,极枯降水年份的泉水流量2.727m3·s-1作为百泉泉域裂隙岩溶地下水开采资源量,可在平水年和丰水年的丰水期有泉水自然涌出,重现自然美景,保护风景泉水旅游资源,做到先观后用,达到保泉和供水的目的。 相似文献
6.
7.
为了解决工程造价预测的时效性问题,针对传统线性时间序列预测模型可靠性不高的缺点,引入混沌相空间重构和支持向量机技术,并将两者耦合组成一种非线性预测模型,再利用ARIMA在整体线性趋势预测方面的优越性,对非线性模型进行修正。混沌SVM和ARIMA预测构成组合模型的两个子过程,将两个子过程的预测结果综合平均即可得到最终预测结果。经实例计算,组合模型比最大Lyapunov指数、ARIMA和只将相空间重构与SVM进行耦合的方法拟合效果好,预测精度高,证明其的确具有线性趋势拟合和非线性波动拟合的双优势。 相似文献
8.
9.
相空间重构的支持向量机预测模型应用十分广泛,在城市供水量预测方面也占据着重要地位,传统的预测模型趋向于将重构的相空间整体带入,这样可能存在引入无效相点从而影响预测精度的问题,基于此将演化追踪法引入相空间重构的预测模型对有效相点进行筛选,优化预测模型的训练样本,达到提高预测精度目的。利用MATLAB编程软件将演化追踪法用于城市供水量的预测,预测结果的平均绝对误差由0.52%降低到了0.29%,证明了演化追踪法的可利用性与有效性。 相似文献
10.
影响空调负荷因素比较多,且难于确定和提取,这就造成空调负荷的拟合和预测精度较低.在对空调负荷时间序列混沌特性分析的基础上,利用嵌人相空间来确定前期影响因子,建立了基于混沌相空间技术的BP神经网络模型.模型既能考虑到影响空调负荷时间序列的动力因子,又能解决网络输入单元数确定的困难,并能利用神经网络超强的非线性映射功能,结合空调负荷实例的拟合与预测,表明其结果合理,预测精度较高. 相似文献
11.
晋陕峡谷地区岩溶地下水的同位素及水化学分析 总被引:1,自引:0,他引:1
本文研究采用同位素和水化学分析的方法对本地区 2个泉域即天桥和柳林泉域的地下水系统进行分析。研究发现 :(1)天桥泉域北部大部分地区岩溶地下水主要接受了大气降水的补给 ,同时在局部地区也接受了地表水的补给 ;天桥泉域南部岩溶地下水接受了不同高程的大气降水的补给 ,迳流较为活跃。 (2 )柳林泉域岩溶水系统存在浅部和深部 2个子系统 ,来自这 2个系统的地下水互相混合 ,形成了排泄区较为复杂的地下水化学类型 相似文献
12.
Upper Jurassic karstified limestones in southern Germany are characterized by a dual-porosity system. Here we present a new model that describes the basic processes from recharge to discharge. While recharge is distributed to both main conduits and fractures, only minor amounts of the total runoff reach the karst system directly (direct runoff) while most water infiltrates into the fractured matrix. This mechanism expels pre-event water from the fissured matrix and into karst conduits. Additionally, matrix storage exponentially decreases during baseflow conditions. Altogether, direct runoff, pre-event water and baseflow fill the karst system and make up the spring discharge. Based on a time series of 40 years of recharge data, the model simulates daily changes of groundwater volumes within the two storage systems. Prerequisites are recession coefficients for baseflow and conduits. A distribution coefficient calculates the amount of direct runoff and a second coefficient the amount of pre-event water. Calculated discharges were in good agreement with the spring hydrograph. 相似文献
13.
Documenting and understanding water balances in a karst watershed in which groundwater and surface water resources are strongly interconnected are important aspects for managing regional water resources. Assessing water balances in karst watersheds can be difficult, however, because karst watersheds are so very strongly affected by groundwater flows through solution conduits that are often connected to one or more sinkholes. In this paper we develop a mathematical model to approximate sinkhole porosity from discharge at a downstream spring. The model represents a combination of a traditional linear reservoir model with turbulent hydrodynamics in the solution conduit connecting the downstream spring with the upstream sinkhole, which allows for the simulation of spring discharges and estimation of sinkhole porosity. Noting that spring discharge is an integral of all aspects of water storage and flow, it is mainly dependent on the behavior of the karst aquifer as a whole and can be adequately simulated using the analytical model described in this paper. The model is advantageous in that it obviates the need for a sophisticated numerical model that is much more costly to calibrate and operate. The model is demonstrated using the St. Marks River Watershed in northwestern Florida. 相似文献
14.
Climate variability induces considerable interannual fluctuations in spring discharge, especially in mountain areas, where groundwater is recharged mainly by rain and snow melt. This study presents the discharge climatological model (DISCLIM), which was developed to test a complexity‐reduced approach to perform historical reconstruction in the lack of physical assumptions. The Mount Cervialto aquifer (Southern Italy) is the test site, where a powerful karst spring is monitored since the 1920s and is very sensitive to climatic conditions. DISCLIM incorporates seasonal precipitation and climate indicators only. Despite its simplicity, DISCLIM has been able to well estimate the yearly fluctuations of discharge hydrological, explaining about 90% of the interannual variability at the calibration stage, and more than 80% at validation stage. This means that DISCLIM can be easily used for estimating the discharge in historical times, when no all the hydrological balance data are available for the purpose. 相似文献
15.
Ulf Mohrlok 《Grundwasser》2014,19(1):73-85
Groundwater flow in karst systems is mainly determined by the efficacy of the drainage through karst conduits. A great challenge for the numerical modelling of such groundwater flow consists in the lack of knowledge about this drainage system. Here, an approach is presented deriving the conduit network for the spring catchment Gallusquelle, central Swabian Alb, from different geographic information. Using this conduit network the transient groundwater flow in the catchment was simulated numerically applying the software ROCKFLOW (Wollrath & Helmig, SM-2 Strömungsmodell für inkompressible Fluide, Theorie und Benutzerhandbuch. Technischer Bericht Institut für Strömungsmechanik, Universität Hannover, 1991). The influence of different parameters on the hydraulic behaviour of the karst system could be demonstrated in a parameter study by means of simulated spring discharge and groundwater tables. In particular, the superposition of drainage behaviour and intermediate storage of groundwater from the conduits in the matrix along with recharge events could be derived in dependence of the varied parameters. 相似文献
16.
17.
基于水动力场控制的岩溶塌陷预警预报研究 总被引:2,自引:0,他引:2
岩溶塌陷是山东省近年来濒发的主要地质灾害之一,主要分布于泰安、枣庄、临沂、莱芜等覆盖型岩溶发育区,其发展、发生主要受岩溶水不合理开采所致,水位是影响岩溶塌陷的重要因素。本文以临沂市城区的岩溶塌陷为例来探讨岩溶塌陷预警预报思路,首先是在完善的监测网络基础上,建立预警模型,通过水文地质模型预测可行的地下水开采方案,并模拟城区一带的岩溶塌陷区水位的变化趋势,人为控制地下水动力场,进而建立岩溶塌陷动力预警预报系统,可为减少或避免该灾害的发生及其防治提供科学依据。 相似文献
18.
对国内外应用较广的公路交通量预测方法——"四阶段"法进行概述,总结交通量预测的影响因素。并结合某公路工程可行性研究报告,利用BP神经网络对影响因素及预测结果进行敏感性分析,找出影响交通量预测工作的最敏感因素。为今后的预测工作提供思路和帮助,从而提高预测结果的准确性。 相似文献