应用于水文预报的优化BP神经网络研究

利用广东省滨江流域的水文观测资料，建立了以前期降水量为预报因子、以水位为输出的BP人工神经网络水文预报模型。首先采用了合理的方法进行样本组织，进而利用最优子集回归技术进行输入因子的确定，然后进行了不同隐层节点数、不同转移函数、不同训练算法的组合试验，确定了应用于水文预报中的优化BP神经网络：网络结构为8-9-1；转移函数的组合方式为tansig-线性函数；训练算法为采用evenberg-Marquardt(Lm)算法。为便于精度分析，还采用了最优子集回归模型作了研究。结果表明，优化BP网络模型无论在拟合精度还是在预测精度上都高于最优子集模型。总的来说BP网络是一种精度较高的水文预测模型。     

石嘴山市浅层地下水的环境背景值

根据水文地质条件将研究区划分为单一潜水区和多层结构区两个地下水环境背景值单元,以石嘴山市大武口区和惠农区1997年和2008年的浅层地下水水质分析数据为依据,主要采用迭代标准差法和计算分布函数法对两个背景值单元内的15项指标进行了背景值范围的计算.在此基础上比较分析了研究区背景值的时空分异特征,结果表明,气象、水文、人类活动等变化较大的因素会使地下水的化学组分随时间呈现出一定的变化,即地下水环境背景值的时间差异性;而地形地貌、含水层岩性、地下水径流条件的差异是影响不同地下水环境背景值单元内相同指标背景值差异的主要因素,即背景值具有空间差异性.对地下水环境背景值进行水质评价,结果表明,2008年较1997年,地下水质量变差.     

银川地区地下水水化学特征演化规律及水质评价

研究地下水的水化学特征演化规律及水质评价,对地下水水资源保护和可持续开发利用具有重要意义.本文根据银川地区1991—2016年地下水水质监测资料,运用经典统计学、Piper三线图解法、Gibbs图解法、离子比值法对研究区地下水水化学特征及演化规律进行分析,并利用模糊综合评价法对研究区62个水样点进行水质评价.结果表明,26年来研究区地下水化学组分未发生明显变化,为偏碱性硬水.多年地下水中各离子含量排序一致,阳离子含量排序为:Na~+Mg~(2+)Ca~(2+)K~+,阴离子含量排序为HCO~-_3 Cl~-SO■.当前银川地区地下水水质基本满足生活饮用水的标准,山前冲洪积平原地区水质良好,丰登镇及北部地区水质最差,部分井孔的TDS大于2000 mg·L~(-1).地下水水质在1996年最差,自2006年以来得到明显改善.研究区地下水化学组分及水质受到水文地质条件、蒸发浓缩作用、岩石风化作用、补给水成分及人类活动的综合影响.     

大安市苏打碱土区地下水质与灌溉盐渍化风险评价

测试了吉林省大安市苏打碱土区6—7、20—30和60—70m^3种深度地下水的基本化学特征，对利用地下水灌溉可能引起的土壤盐化、碱化和盐碱化问题及其灌溉适宜性进行了评价。结果表明，各深度地下水均具有不同程度的苏打化特征，地下水中普遍含有较多NaHCO3；30m深度以内地下水中含有较多HCO3^-和Cl^-，用其作为灌溉水源存在潜在的盐化和碱化双重威胁；从各盐化和碱化指标看，该区60m深度以下的地下水质较好，作为灌溉用水不会产生盐渍化风险。30m深度以内的地下水均不适合灌溉。     

高含硅量高盐地下水处理工艺

本文通过对高含硅量，高盐地下水这一特殊水质进行纯化处理研究，经实际运行及分析，得出一套行之有效的工艺流程。     

贵州高原喀斯特流域浅层地下水化学特征及质量评价——以普定后寨河为例

为研究贵州典型喀斯特流域浅层地下水化学特征及水质状况,于2015年5月—12月对贵州普定后寨河流域进行现场调查并采集浅层地下水样品28个,测定K~+、Na~+、Ca~(2+)、Mg~(2+)、HCO_3~-、NO_3~--N、NH_4~+-N等20种指标,利用Duncan差异显著性检验法进行水化学特征分析,用水质质量综合评价法对水质进行评价.结果表明,浅层地下水呈微碱性,主要阴离子为HCO_3~-、SO_4~(2-)、Cl~-,主要阳离子为Ca~(2+)、Mg~(2+).从季节变化来看,Cl~-、SO_4~(2-)、K~+、TN是春季夏季秋季,在冬季时其含量升高,HCO_3~-是春季≈夏季秋季冬季.NO_3~--N是夏季春季秋季冬季.Na~+、Mg~(2+)是秋季和冬季较低.pH、NH_4~+-N是春季和秋季较低,4个季节中呈"N"型分布,可见,在喀斯特山区浅层地下水具有明显的季节变化特征.水质综合评价表明,在喀斯特山区秋、冬季节水质较春、夏季节好,总体质量较好.而从单项组分来看,部分地区有的指标已超过Ⅳ类水的限值,主要是总Fe和NH_4~+-N质量浓度较高影响水质.总Fe的最高质量浓度为1.2948 mg·L~(-1),NH_4~+-N的最高质量浓度为0.71 mg·L~(-1),这与当地人类生产生活活动有关,对浅层地下水质量具有潜在的影响.综上,研究贵州高原喀斯特流域浅层地下水化学特征及水质,可为喀斯特山区水资源的保护和管理提供科学依据.     

重庆主城都市区地下水质量与污染评价——基于层级阶梯评价法

为查明重庆主城都市区的水质及污染状况并区分天然背景和人类污染对地下水水质的影响,进行了地下水采样和测试,利用层级阶梯评价法,对地下水水质和污染状态进行了分析与评价。结果表明,研究区内地下水受到天然背景和人类污染的共同影响,Ⅲ类和Ⅳ类水为主要水质类别,各占45.9%,主要的超标指标为硝酸盐、锰、砷、总硬度等,研究区地下水有机污染程度为轻微,农业面源污染和未经处理的生活排污是地下水污染的主要原因。相比传统的单因子评价数学方法,层级阶样评价法将污染源纳入考虑,突出了人类污染属性,对于地下水污染防治可以提供更有针对性的决策。     

廊坊市区地下水质量现状评价与分析

为阐明人类活动影响下廊坊市区地下水环境状况,2015年在研究区进行现场调查采集并分析了深浅层(包含3组地表水)水样50组,测试指标包括有机指标(30项)和无机指标(31项).利用基于分类指标的单指标综合评价法进行本区地下水质量评价.结果表明,研究区地下水质量综合评价结果较差,深层地下水相比浅层地下水水质较好;从分类指标来看,对区域地下水质量影响最大的为无机毒理指标,其次为一般化学指标,毒性(类)金属指标及有机指标对地下水质量影响很小,各项指标主要影响浅层地下水;单项组分分析表明,深、浅层地下水水质贡献指标相差较大,其中影响浅层地下水水质的主要指标有总硬度、"三氮"、砷、氟化物、铁锰等,深层地下水影响指标主要为氟化物(50.0%),其次为砷(13.6%);有机组分少量指标检出但无超标.分析认为原生水文地质环境和人类活动是影响该区地下水质量的主要因素,是导致地下水中污染组分异常的根源.     

石油类污染场地地下水抽水过程水质变化成因

在某石油类污染场地抽水试验过程中对地下水五项常规水质指标进行了连续监测.通过对指标的相关性和主成分分析,基于抽水、弥散试验,揭示了该场地抽水过程中水质变化的原因,探讨了以水质指标描述石油类污染场地地下水净化的方法.结果显示,水质变化是由昼夜水温变化、抽水和生化耗氧作用三者引起.其中,水温变化是由昼夜变化引起;电导率减小...     

华北平原地下水有机污染特征初步分析

各种污染造成地下水中有害物质逐年增加,地下水水质逐渐下降,严重危及了饮用水安全。针对此情况,按照相关的要求,采集了245组水样。样品由有资质的测试单位测定。依据对地下水质调查的初步分析,华北平原地下水中已有机物检出。对所有水样测试结果的统计显示：检出的有机污染物共有32种。检出率大于4.5%的有氯仿、甲苯、四氯乙烯、苯并[a]芘、氯苯、苯和1,2-二氯苯,其中个别水样中的四氯化碳、苯并[a]芘和挥发性酚类三项超过了饮用水标准（据地下水质量标准报批稿,中国地质调查局,2007）,超标率分别为1.22%、2.45%和1.90%。个别在深层地下水样品中也检出了微量有机污染物,这可能与取样井的混合开采、止水不佳有关。从区域上来看,在人类活动相对强烈的城镇地区地下水中,有机物检出率相对较高,特别是在渗透性好的山前地带,污染物较易进入到地下水中。通过分析发现,地下水有机物的高检出率与地下水硬度的升高有一定相关性。由于地下水有机污染在时间和空间上差异一般较大,建议对已发现的异常点进行重复取样测试,积累水质数据,以利于将来的综合分析。     

欧盟土壤环境评价监测项目及其对我国农用地质量监测的启示

＂十二五＂期间,我国国土资源调查评价的重点任务之一是建立农用地质量监测体系。当前我国农用地质量变化监测技术仍然处于研究探索阶段,欧盟土壤环境评价监测项目是较为成熟的监测项目,可为中国农用地质量监测工作提供借鉴。文章分别就欧盟土壤环境评价监测的相关研究和国内学者对农用地质量监测的研究进行了综述,认为欧盟土壤环境评价监测项目在指标选取、监测网络建立、数据更新技术等方面建立了较为成功的机制。未来中国农用地质量监测工作可在关注土壤环境质量、分类土壤退化原因、针对性地选取指标、应用DPSIR模型、确定监测点的方法、充分利用原有监测网络和监测数据等方面借鉴欧盟经验。     

神经网络模型在辽河水质量评价中的应用

为准确的反映辽河水质量状况和发展趋势，在对辽河水质监测采样和实验分析的基础上，建立了水质评价的BP神经网络模型，应用该模型对辽河的水质现状进行了质量综合评价，4个监测断面的评价结果均在4～4．5之间，水污染情况较为严重。该评价结果与监测断面污染项目的超标情况分析的结论吻合，评价效果较好。     

Health risk of arsenic in the alluvial aquifers of Lahore and Raiwind,Punjab Province,Pakistan: an investigation for safer well water

Environmental pollution has generated release of high amounts of arsenic (As) which ultimately are detected in the water of Indus Basin Punjab, Pakistan. The area is characterized by a semiarid climate, and alluvial deposits. This investigation, an extension of previously reported As-affected area in Lahore and Kasur, aimed to (1) assess the extent of water contamination in this area and (2) determine possible safer sites for future water use. In a comparative study, total As contamination of underground water from individual dwellings and community water supply of some villages located at the boundary of the Sheikhopura and Lahore districts, Pakistan were measured to compare with previously published data of villages located at the boundary of the Lahore and Kasur districts. The results showed variable levels of As in shallow drinking water wells and average concentration exceeding WHO guidelines value. As levels ranged from below 5.2 to 80 µg/L and mean 45.5 µg/L. The As concentrations were higher than WHO limits but lower than previous studied area. In addition, high salinity was found to be a serious concern for deteriorating groundwater quality rendering it unsuitable for drinking. Groundwater is predominantly of the Na–HCO₃ type with slightly alkaline pH. High pH values and competition of As with HCO₃ may serve as an important process for mobilization of As in the shallow groundwater of the region. Continuous monitoring and expansion of monitoring systems are necessary to establish safer wells within As-contaminated areas.     

基于BP神经网络算法的密云水库水质参数反演研究

密云水库是北京市重要的地表饮用水源地,在保障首都水源安全方面起着重要作用,而密云水库水质参数的区域监测已成为当前亟待解决的问题。为了实现密云水库水质参数大范围、实时获取,该文基于遥感和GIS技术,采用BP神经网络算法,结合地面监测数据和Landsat 8遥感影像,分别建立了反演总磷、总氮、氨氮和COD(化学需氧量)4个水质参数的BP神经网络模型,并反演了密云水库2013-2018年非结冰期主要水质参数,分析了密云水库主要水质参数的年际变化特征、季节变化特征和空间分异特征。结果表明,(1)水质参数的Landsat 8敏感波段分别为:总氮为1、4波段,氨氮为1-7波段,总磷为1、3-7波段,COD为2-5波段。(2)密云水库主要水质参数在2013-2018年总体呈下降趋势,氨氮和COD为Ⅰ类水质,总磷为Ⅱ类水质,总氮为Ⅲ类水质。(3)4个水质参数指标春季最高、秋季次之、夏季最低,总氮、总磷、氨氮和COD的春季值分别是夏季值的1.08、1.36、1.6、1.45倍。(4)密云水库不同水质参数的空间差异性较大,总体来看,水库北部和东部的4个水质参数含量相对较高,这种分布与北部和东部村庄密集以及密云水库两大入库河流有关。综上所述,基于BP神经网络算法的密云水库水质反演研究是可行的,且得到了较为可信的研究结果,该研究可为密云水库水质管理与政策制定提供重要的科学依据。     

Design of a water quality monitoring network in a large river system using the genetic algorithm

Despite several decades of operations and the increasing importance of water quality monitoring networks, the authorities still rely on experiential insights and subjective judgments in siting water quality monitoring stations. This study proposes an integrated technique which uses a genetic algorithm (GA) and a geographic information system (GIS) for the design of an effective water quality monitoring network in a large river system. In order to develop a design scheme, planning objectives were identified for water quality monitoring networks and corresponding fitness functions were defined using linear combinations of five selection criteria that are critical for developing a monitoring system. The criteria include the representativeness of a river system, compliance with water quality standards, supervision of water use, surveillance of pollution sources and examination of water quality changes. The fitness levels were obtained through a series of calculations of the fitness functions using GIS data. A sensitivity analysis was performed for major parameters such as the numbers of generations, population sizes and probability of crossover and mutation, in order to determine a good fitness level and convergence for optimum solutions. The proposed methodology was applied to the design of water quality monitoring networks in the Nakdong River system, in Korea. The results showed that only 35 out of 110 stations currently in operation coincide with those in the new network design, therefore indicating that the effectiveness of the current monitoring network should be carefully re-examined. From this study, it was concluded that the proposed methodology could be a useful decision support tool for the optimized design of water quality monitoring networks.     

An emerging market for groundwater remediation in China: Policies,statistics, and future outlook

There is a rapidly emerging and potentially huge market for the remediation of contaminated groundwater in China. The Chinese government published a Water Action Plan in April 2015, a Soil Action Plan in May 2016, and a draft Soil Pollution Prevention and Control Law in June 2017. All of these new policies and regulations put pressures on local governments and contaminated site owners, obliging them to conduct site investigation and to cleanup contaminated groundwater. The Chinese population in northern regions heavily depend on groundwater, with nearly 70% of water supply coming from aquifer sources in the Beijing-Tianjin-Hebei region. However, poor groundwater quality due to natural geochemical background and anthropogeic pollution is a serious concern, with poor or very poor quality water observed in nearly 80% of groundwater monitoring wells in 17 northern provinces. Shallow groundwater in many areas has been contaminated by toxic pollutants such as heavy metals and chlorinated organic compounds. There is an urgent need to better understand the situation and to conduct groundwater remediation at contaminated sites. The Chinese government is investing heavily in the research and development for groundwater remediation, which is expected to greatly add to the quality and quantity of groundwater remediation projects in the near future.     

城市给水管网余氯的预测模型

选择余氯为研究对象,以南方某市给水管网水质监测的数据为基础,使用线性回归和非线性神经网络(ANN)方法建立模型,找到了一种利用在线监测数据和人工监测数据实时预测管网余氯的方法。通过建立给水管网水质模型,可以由监测系统动态回传的数据来实时的预测下一天人工点的水质。模拟的结果显示ANN模型比线性回归模型有更好的预测能力,预测的平均相对误差:ANN模型为14.9%,线性回归模型为25.8%。使用ANN模型可以实现实时预测。     

A novel hybrid model for water quality prediction based on VMD and IGOA optimized for LSTM

● A novel VMD-IGOA-LSTM model has proposed for the prediction of water quality. ● Improved model quickly converges to the global optimal fitness and remains stable. ● The prediction accuracy of water quality parameters is significantly improved. Water quality prediction is vital for solving water pollution and protecting the water environment. In terms of the characteristics of nonlinearity, instability, and randomness of water quality parameters, a short-term water quality prediction model was proposed based on variational mode decomposition (VMD) and improved grasshopper optimization algorithm (IGOA), so as to optimize long short-term memory neural network (LSTM). First, VMD was adopted to decompose the water quality data into a series of relatively stable components, with the aim to reduce the instability of the original data and increase the predictability, then each component was input into the IGOA-LSTM model for prediction. Finally, each component was added to obtain the predicted values. In this study, the monitoring data from Dayangzhou Station and Shengmi Station of the Ganjiang River was used for training and prediction. The experimental results showed that the prediction accuracy of the VMD-IGOA-LSTM model proposed was higher than that of the integrated model of Ensemble Empirical Mode Decomposition (EEMD), the integrated model of Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Nonlinear Autoregressive Network with Exogenous Inputs (NARX), Recurrent Neural Network (RNN), as well as other models, showing better performance in short-term prediction. The current study will provide a reliable solution for water quality prediction studies in other areas.     

BP模型的改进及其在大气污染预报中的应用

针对传统BP模型存在着训练速度较慢、局部极值以及最佳网络结构无法准确确定的不足，进行了改进，应用于城市空气污染预报，建立大气污染浓度的神经网络预测模型。计算结果表明，应用改进的BP模型进行大气污染预报能够得到更好的预测结果，具有很强的实用性。