滇池水污染物浓度预测的人工神经网络模型

本文针对滇池日益严重的水污染现状,根据云南昆明西苑隧道断面2004年-2010年的监测资料,建立了基于BP神经网络的主要污染指标预测模型,并对其进行训练检验,研究结果表明：独立样本中pH、溶解氧、氨氮、高锰酸盐浓度的预测值与监测值的线性相关系数分别为0．952、0．967、0．945、0．936。结果证明该模型预测精度满足要求,通过准确地预测湖泊水污染物可以为治理湖泊营养化和综合利用水资源、规划管理、决策提供重要的科学依据。     

三度空间视角下的装配式建筑施工安全绩效评价*

为了解决现有的装配式建筑施工安全绩效评价已不能满足建筑业信息化背景下发展需求的问题,基于“三度空间”视角,从物理、社会、信息3个层面,构建包含6个一级指标以及23个二级指标的装配式建筑施工安全绩效评价体系;阐述各指标之间的相互关联;采用ANP网络模型对各指标赋权;利用灰色聚类法对施工安全绩效指标进行评价。研究结果表明:该评价模型能够分析装配式建筑在社会、物理和信息3个空间维度的安全绩效,得出装配式建筑项目的整体安全绩效水平并识别出关键控制指标,可为装配式建筑安全绩效评价提供新的思路。     

医院智能应急疏散导航方法与物联网应用研究*

为解决大型复杂医院建筑紧急疏散下的路线指引问题,提出适合复杂空间结构的疏散拓扑地图自动生成方法,同时提出基于组合神经网络的疏散决策模型,以应对医院建筑中各种静态动态的危险源,基于该疏散指引方法,研发智能疏散导航物联网装置Nav-IoT。研究结果表明:该方法可实时采集医院多种危险情况,通过决策网络和多种物联网智能设备进行动态决策,并发出明确移动指引信号,能够有效提升疏散过程的安全度和效率,满足大型复杂医院建筑内的智能化、动态化的人员应急疏散需求。     

长江水环境质量监测网络运行体系初步构建

为精准、客观、全面评价长江水环境质量和支撑长江大保护,通过研究制定一系列管理制度和技术规范,以优化调整的水环境质量监测网络为基础,建立了涉及管理制度、监测技术、质量控制和数据审核等方面的水质自动监测与采测分离手工监测相结合的网络运行机制。研究自动与手工监测的总磷前处理方式,实现了自动与手工监测数据的匹配和融合,并将融合数据应用于长江水环境质量评价、预警、考核与排名,实现了长江水环境质量监测网络的业务化运行。     

Water-lifting and aeration system improves water quality of drinking water reservoirs: Biological mechanism and field application

Reservoirs have been served as the major source of drinking water for dozens of years. The water quality safety of large and medium reservoirs increasingly becomes the focus of public concern. Field test has proved that water-lifting and aeration system (WLAS) is a piece of effective equipment for in situ control and improvement of water quality. However, its intrinsic bioremediation mechanism, especially for nitrogen removal, still lacks in-depth investigation. Hence, the dynamic changes in water quality parameters, carbon source metabolism, species compositions and co-occurrence patterns of microbial communities were systematically studied in Jinpen Reservoir within a whole WLAS running cycle. The WLAS operation could efficiently reduce organic carbon (19.77%), nitrogen (21.55%) and phosphorus (65.60%), respectively. Biolog analysis revealed that the microbial metabolic capacities were enhanced via WLAS operation, especially in bottom water. High-throughput sequencing demonstrated that WLAS operation altered the diversity and distributions of microbial communities in the source water. The most dominant genus accountable for aerobic denitrification was identified as Dechloromonas. Furthermore, network analysis revealed that microorganisms interacted more closely through WLAS operation. Oxidation-reduction potential (ORP) and total nitrogen (TN) were regarded as the two main physicochemical parameters influencing microbial community structures, as confirmed by redundancy analysis (RDA) and Mantel test. Overall, the results will provide a scientific basis and an effective way for strengthening the in-situ bioremediation of micro-polluted source water.     

A Comparison of the Precipitation Chemistry Measurements Obtained by the CAPMoN and NADP/NTN Networks

Precipitation chemistry measurementsobtained by the Canadian Air and PrecipitationMonitoring Network (CAPMoN) and the U.S. NationalAtmospheric Deposition Program/National Trends Network(NADP/NTN) have been examined using more than 7 yrof collocated data from two sites, namely, Sutton,Quebec, Canada and State College, Pennsylvania, U.S.A.In the case of the CAPMoN data, weeklyprecipitation-weighted mean concentrations, totalsample depths and total standard gauge depths werecomputed from daily data and compared to thecorresponding weekly sampling data of the NADP/NTNnetwork. Seasonal and annual precipitation-weightedmean concentrations and deposition values were alsocomputed for both networks and compared. Statisticallysignificant between-network biases were found to existin the weekly results for most of the measuredvariables, particularly standard gauge depth, sampledepth, pH, H⁺, NO₃ ^-,NH₄ ⁺,Na⁺; the NADP/NTN values were consistently lowerthan those of CAPMoN with the exception of pH andNa⁺. The magnitude of the biases was less than35% of the median CAPMoN weekly value for the 7 yr. For most of the measured parameters, thevariability of the between-network differencesrepresented less than 20% of the median CAPMoN weeklyvalue. Both the between-network biases andvariabilities were functions of several physicalparameters, the most dominant being the sample depthand the ionic concentration. For seasonal and annualdeposition values, statistically significantbetween-network biases were found for H⁺,SO₄ ^2-, NO₃ ^-,Ca²⁺,NH₄ ⁺ for both periods; for Mg²⁺ andK⁺ for seasonal data; and Cl^- for yearlydata, with the NADP/NTN deposition values being lowerthan those of CAPMoN. The relative biases ranged from7 to 37%. Part of the between-network bias in thedeposition estimates was directly attributable to astrong bias in the standard gauge depths of the two networks.     

建立有效回收网络,防治危险废物污染环境

工业废弃物中，废酸碱、重金属、废矿物油等都属于危险废物，我们日常生活中的废电池、灯管、显像管、空调、冰箱等都含有大量成分的危险废物。开展危险废物的有效回收，建立有效回收网络，才能从根本上防治危险废物进入环境，防治危险废物给人类与环境造成严重灾难。     

大别山区经济发展的交通运输条件分析

:2 0世纪 90年年代 ,大别山区的社会经济有了相当大的发展 ,交通运输网络已初步形成 ,整个内山地区也被铁路和高速公路等交通大动脉所环抱 ,然而经济发展水平和发展速度还无法与周边地区相比。从分析大别山交通现状着手 ,认为大别山区经济发展滞后的主要制约因素之一是区内交通的落后。指出由于区位、地形、社会经济发展水平及行政隶属上的限制 ,大别山区内部出山通道少 ,公路路况差 ,路网连接率低 ,缺少综合运输网络 ,使得内山地区很少受到交通大动脉的影响 ,正如一三角阴影区覆盖其上。然后提出了解决问题的策略 :其一是改善内山公路路况 ;其二是加强内部协作 ,提高内山公路联接率 ;第三是建立综合运输网络 ;最后是改变交通运输投资和管理体制。     

基于EOF和BP网络的库群径流预测

在考虑多个水库径流相关性的基础上，基于自然正交函数，将多维径流序列分解成时不变的空间特征和随时间变化的权重系数矩阵，并运用网络对时间权重系数变化规律进行分析，从而对库群径流做出预测，以某区域中的三座中型水库为例，选用灰关联模型进行比较预测，取得了较好的结果。     

基于小波网络的旅游地复合系统协调开发评价与实证

本文在阐明旅游地复合系统（简称TDCS）及其协调开发内涵的基础上，利用小波网络 的非线性特性对旅游地复合系统协调开发进行评价，避免了过去只评价旅游地开发状态过于笼统，线性规律不符合实际情况的弊病。建立了基于小波网络的旅游地复合系统协调开发评价模型，并以秦皇岛为例进行了实证研究，为旅游地开发管理提供了一种新的思路、新的方法。