共查询到20条相似文献，搜索用时 399 毫秒
1.
多层砖房在遭遇地震时破坏较为严重,会给人民的生命与财产造成巨大的损失。将神经网络与遗传算法相结合,建立遗传神经网络对多层砖房的震害进行了预测。样本的预测结果与实际结果一致,说明应用遗传神经网络算法对多层砖房的震害进行预测是可行。 相似文献
2.
土层参数随机性对多层土模型传递函数的影响 总被引：1，自引：0，他引：1
土体的地震反应与结构震害密切相关,是地震工程中的一个重要问题,也是抗震设计的基础之一.然而目前的工程场地地震反应分析中存在着诸多不确定性因素,如何估计这些不确定的影响是一个亟待解决的问题.本文采用泰勒级数展开估计土层参数随机性对场地传递函数的影响.以两层土模型为例,对多层土模型在土层参数服从独立高斯分布假定下进行公式推... 相似文献
3.
《河北农业大学学报》2015,(6)
以FNN算法实现震害预测,确定预测因素和指标隶属度,建立模糊BP模型,进行神经网络构建,以实地测量的方式,采集建筑样本进行试验,分别研究了各个因素对建筑震害的影响,包括高宽比、层数、横向跨数、纵向跨数、地震烈度、抗震横墙数的影响,对建筑震害主导因子进行确定,以优化BP网络进行建筑震害估算,仿真结果证实:估算效果较好。 相似文献
4.
研究水中溶解氧含量的预测问题,可为水厂生产、水产养殖业、地表水环境的管理提供科学依据。影响水中溶解氧的各因素关系复杂,很难进行数学建模,神经网络由于具有较强的非线性问题处理能力而被广泛应用于水中溶解氧的预测研究。由于传统神经网络具有陷入局部极小、收敛速度慢并且网络冗余度较大等缺陷,为提高网络的正确预测能力和泛化能力,采用多层激励函数(modulation transfer function,简称MTF)与量子神经网络(quantum neural network,简称QNN)相结合建立水中溶解氧预测模型,但若用神经网络作为前馈型网络,它的沿梯度下降的算法会使其在进行网络训练时有可能陷入局部极小,针对这一缺陷,提出一种可以使量子神经网络逃离局部极小点的算法,并将它应用到量子神经网络中进行网络误差调整。仿真结果表明,与传统BP(back propagation)神经网络预测结果相比,采用多层激励函数的量子神经网络(modulation transfer function-quantum neural network,简称MTF-QNN)预测水中溶解氧含量的预测精度更高。 相似文献
5.
基于人工神经网络的粮食产量预测模型 总被引：7，自引：0，他引：7
研究了人工神经网络在经济预测中的应用问题,探讨利用人工神经网络进行农业粮食产量预测的方法。提出一种基于多层前馈BP神经网络的农业粮食产量预测模型,可以得到影响粮食产量的主要因子和粮食产量之间的非线性映射关系。并通过实例验证了神经网络模型的预测精度明显高于线性回归模型的预测精度。 相似文献
6.
基于模糊聚类的神经网络虫情预测 总被引：1，自引：0，他引：1
1基于模糊聚类的神经网络预测
首先采用模糊聚类对所有样本进行预处理,再把去噪后的数据作为神经网络的输人数据进行训练和预测。
1．1基于模糊聚类的神经网络结构3层BP神经网络具有令人满意的对连续映射的逼近能力,可以满足预测的要求,因此,采用3层BP神经网络作为研究模型。3层BP神经网络由输入层、隐含层和输出层组成。 相似文献
7.
针对传统方法预测网络流量精度较低的问题,提出了一种基于改进双线性递归神经网络模型(BLRNN)的非线性网络流量预测方法.首先,给出了双线性递归神经网络的定义及网络结构描述;然后从网络结构和修剪过程两方面对双线性递归神经网络进行优化改进,采用遗传算法全局搜索进行修剪;最后,通过真实工况下的网络流量数据用仿真试验对模型性能进行分析.试验结果表明,优化后双线性递归神经网络模型大幅降低了算法复杂度,提高了计算效率,与传统的多层神经网络预测方法相比,该方法预测精度更高.同时,该方法也为其他具有相似特征的非线性预测问题提供了一种新的研究思路. 相似文献
8.
影响土坝震害因素的探讨 总被引：1，自引：0，他引：1
黄代清 《云南农业大学学报(自然科学版)》1991,6(2):126-128
《地震工程研究报告集》第四集,"土坝的震害规律和抗震设计问题"一文认为影响土坝震害的主要因素是坝高和震中距,并以通海地震、海城地震等116座土坝的震害调查,作出附图。文中写道:根据附图的资料可以发现如下的趋势,对于7至8级的地震,当震中距小于20km时,如果坝高 H>5+1.67L (1) 时,土坝一般不会产生严重震害,最可能产生一般震害。 相似文献
9.
《浙江大学学报(农业与生命科学版)》2017,(1)
在南美白对虾高密度、规模化围塘养殖生产中,水质参数的监测、评价及预警是至关重要的。以杭州市某南美白对虾基地日常养殖水质为研究对象,选取温度、pH值、溶解氧、氧化还原电位等4项指标作为预测参数,建立拓扑结构为40-14-4的3层前馈反向传输(back propagation,BP)神经网络模型,即以连续10个时间单位的预测变量为输入层,隐含层节点数为14个,输出层变量为温度、pH值、溶解氧和氧化还原电位。为克服传统BP神经网络存在的收敛速度慢、易陷入震荡和泛化能力不强等缺点,采用模糊方法优化了自适应变步长BP神经网络算法,缩短了BP神经网络的训练时间,提高了网络收敛效率与稳定度。结果表明,以模糊方法改进的BP神经网络(FABPM)方法具有收敛速度快、预测精度高、稳定度好等特点,对选取的4项水质指标都给出了较好的训练与预测结果,预测的平均相对误差小于2.5%,从而为水产养殖水质预测与评价提供了一种新方法。 相似文献
10.
灌溉用水量的预测对灌区的灌溉管理起着重要的作用。运用神经网络中Elman反馈型神经网络建立了灌溉用水量预测模型,模型输入层神经元数目为4,输出层神经元数目为1,隐含层神经元数目确定采用试验法,最终确定为10。预测结果表明:该方法与传统的预测方法相比,具有网络稳定性高,训练误差曲线比较平滑,模型预报精度较高等优点。 相似文献
11.
刘廉纯 《大连水产学院学报》1996,11(1):28-34
根据建筑抗震设计规范的有关规定，对多层砌体房屋楼层地震剪力的计算。提出了一种经算法，编制了计算程序和表格。通过算例将本文方法与规范方法进行了对比，结果表明本文方法简单、实用、准确。 相似文献
12.
用灰色GM(1,1)模型、BP神经网络模型和证据理论相结合而成的灰色证据神经网络模型,对山东枣庄区域地下水位进行预测。基本思路是运用灰色GM(1,1)模型所得到的预测值按前两年来预测下一年的组合规律分别作为BP神经网络的输入输出;再用BP神经网络输出作为证据理论基本可信度分配函数,使用D-S理论将信息进行二次融合;并用地下水位实测数据对模型进行验证。结果表明,预测模型具有可行性和实用性,为生态农业规划发展提供了科学依据。 相似文献
13.
Ground shaking close to the causative fault of an earthquake is more intense than it was previously believed to be. This raises the possibility that large numbers of buildings and other structures are not sufficiently resistant for the intense levels of shaking that can occur close to the fault. Many structures were built before earthquake codes were adopted; others were built according to codes formulated when less was known about the intensity of near-fault shaking. Although many building types are more resistant than conventional design analyses imply, the margin of safety is difficult to quantify. Many modern structures, such as freeways, have not been subjected to and tested by near-fault shaking in major earthquakes (magnitude 7 or greater). Damage patterns in recent moderate-sized earthquakes occurring in or adjacent to urbanized areas (17), however, indicate that many structures, including some modern ones designed to meet earthquake code requirements, cannot withstand the severe shaking that can occur close to a fault. It is necessary to review the ground motion assumed and the methods utilized in the design of important existing structures and, if necessary, to strengthen or modify the use of structures that are found to be weak. New structures situated close to active faults should be designed on the basis of ground motion estimates greater than those used in the past. The ultimate balance between risk of earthquake losses and cost for both remedial strengthening and improved earthquake-resistant construction must be decided by the public. Scientists and engineers must inform the public about earthquake shaking and its effect on structures. The exposure to damage from seismic shaking is steadily increasing because of continuing urbanization and the increasing complexity of lifeline systems, such as power, water, transportation, and communication systems. In the near future we should expect additional painful examples of the damage potential of moderate-sized earthquakes in urban areas. Over a longer time span, however, we can significantly reduce the risk to life and property from seismic shaking through better land utilization, improved building codes and construction practices, and at least the gradual replacement of poor buildings by more resistant buildings. Progress toward reducing risk from seismic shaking through better building design is slowed by deficiencies in our knowledge about the nature of damaging ground motion and the failure mechanisms in structures. For example, lacking observational data, seismologists must rely on simplified theoretical and numerical models of the earthquake process to estimate near-fault ground motion, especially for earthquakes as large as magnitude 7 and 8. Because such models have not been adequately tested against data, their reliability is unknown. Engineers lack detailed information about failure processes in structures during an earthquake. Although many structures have been instrumented to measure their response to an earthquake, few records have been obtained from buildings that actually sustained significant structural damage and few structures are properly instrumented to measure all the modes of deformation that are likely to contribute to failure. Moreover, the fact that many structures have withstood ground motion more intense than that assumed in their design indicates that conventional methods of design do not take into account important contributions to earthquake resistance by nonstructural elements and by the ability of structural elements to deform inelastically without necessarily causing failure of the structure. It is fortunate when such reserve resistance exists, but better understanding of the sources of reserve strength is needed to determine how large a margin of safety they confer and how they might be affected by changes in construction practices and materials with time. In the next few years we look forward to significant advances in knowledge and to more effective application of what is already known, largely because of substantial funding of research related to seismic engineering by the National Science Foundation (18). The increasing number of strong-motion seismographs operating in seismically active regions (19) will likely provide for the first time a number of records of damaging levels of ground motion. Significant effort is being directed toward obtaining near-fault records, although many probable sites of future large earthquakes remain inadequately instrumented, especially outside the conterminous United States. New and more complete information on building response and damage mechanisms will be obtained by improved instrumentation of structures and through laboratory investigations of failure in structures and structural elements. Further developments in computer technology and in computer modeling techniques will permit more realistic simulations of the seismic response of soils and structures that take into account their inelastic behavior and their strain-dependent properties. Earthquake design codes will continually be revised to better utilize existing knowledge concerning the nature of strong ground motion and the dynamic behavior of buildings during earthquakes and to incorporate new knowledge and also experiences gained from future earthquakes. We believe that application of new knowledge, improvements in earthquake-resistant design and construction, and remedial strengthening or replacement of weak existing structures can significantly reduce our current level of exposure to earthquake hazards. 相似文献
14.
MATLAB在RBF神经网络模型中的应用 总被引：1，自引：0，他引：1
本文介绍了RBF神经网络的基本原理及主要特点,并举例说明了基于MATLAB神经网络工具箱建立RBF神经网络模型及实现仿真的方法. 相似文献
15.
16.
17.
Near-infrared Spectral Detection of the Content of Soybean Fat Acids Based on Genetic Multilayer Feed forward Neural Network 总被引：1，自引：0，他引：1
CHAIYu-hua PANWei NINGHai-long 《东北农业大学学报(英文版)》2005,12(1):74-78
In the paper, a method of building mathematic model employing genetic multilayer feed forward neural network is presented, and the quantitative relationship of chemical measured values and near-infrared spectral data is established. In the paper, quantitative mathematic model related chemical assayed values and near-infrared spectral data is established by means of genetic multilayer feed forward neural network, acquired near-infrared spectral data are taken as input of network with the content of five kinds of fat acids tested from chemical method as output,weight values of multilayer feed forward neural network are trained by genetic algorithms and detection model of neural network of soybean is built. A kind of multilayer feed forward neural network trained by genetic algorithms is designed in the paper. Through experiments, all the related coefficients of five fat acids can approach 0.9 which satisfies the preliminary test of soybean breeding. 相似文献
18.
19.
农业温室二氧化碳模糊控制系统算法 总被引：1，自引：0，他引：1
利用BP神经网络模糊控制器对农业温室二氧化碳进行控制.详细介绍了模糊规则表的生成及神经网络模糊控制器的设计,并使用Mablab进行了仿真学习,仿真结果表明该设计能够有效的对温室二氧化碳进行控制. 相似文献
20.
区域自动气象站测站利于对地震等灾害性的天气进行监控和测试。我们国家在各区域遍布站点,形成区域气象观测站网。以山西省寿阳县气象局的观测站为例,对区域站日常维护、故障原因方面进行探讨,提出常见故障及其修理注意的事项以及日常维护的方法,进而保证区域站运行高效的工作。 相似文献