基于RF-LSTM的鸡舍恶臭气体预测研究

以鸡舍氨气为研究对象,对鸡舍氨气预测模型进行了研究.首先,利用随机森林算法（RF）对影响鸡舍氨气浓度的环境变量进行重要性排序,选取温度、湿度、光照、气象温度、降雨量作为模型的输入变量;在此基础上,构建了基于长短时记忆神经网络（LSTM）的鸡舍氨气浓度预测模型,并将提出的预测模型应用于江苏省宜兴市某养鸡场的氨气浓度预测中,并与LSTM模型、RF-Elman模型和RF-BP模型进行了对比实验,结果表明,基于RF-LSTM模型的预测效果最好,其平均绝对误差（MAE）、平均绝对百分误差（MAPE）和均方根误差（RMSE）分别为0.9183、4.9637%和1.4262;同时,为了验证该模型的性能,本文还实现了不同时间尺度的鸡舍氨气浓度预测,提前2h、3h、4h、5h氨气预测的平均绝对误差（MAE）分别为1.6218、2.1991、2.8553和3.0677.本文提出的预测模型提高了鸡舍氨气浓度的预测精度,可为减少鸡舍恶臭气体排放提供科学依据.     

Performance of different macrophytes in the decontamination of and electricity generation from swine wastewater via an integrated constructed wetland-microbial fuel cell process

Plants constitute a major element of constructed wetlands(CWs).In this study,a coupled system comprising an integrated vertical flow CW(IVCW) and a microbial fuel cell(MFC) for swine wastewater tre atment was developed to research the effects of macrophytes commonly employed in CWs,Canna indica,Acorus calamus,and Ipomoea aquatica,on decontamination and electricity production in the system.Because of the different root types and amounts of oxygen released by the roots,the rates of chemical oxygen demand(COD) and ammonium nitrogen(NH_4~+-N) removal from the swine wastewater differed as well.In the unplanted,Canna indica,Acorus calamus,and Ipomoea aquatica systems,the COD removal rates were 80.20%,88.07%,84.70%,and 82.20%,respectively,and the NH_4~+-N removal rates were 49.96%,75.02%,70.25%,and 68.47%,respectively.The decontamination capability of the Canna indica system was better than those of the other systems.The average output voltages were 520±42,715±20,660±27,and 752±26 mV for the unplanted,Canna indica,Acorus calamus,and Ipomoea aquatica systems,respectively,and the maximum power densities were 0.2230,0.4136,0.3614,and0.4964 W/m~3,respectively.Ipomoea aquatica had the largest effect on bioelectricity generation promotion.In addition,electrochemically active bacteria,Geobacter and Desulfuromonas,were detected in the anodic biofilm by high-throughput sequencing analysis,and Comamonas(Proteobacteria),which is widely found in MFCs,was also detected in the anodic biofilm.These results confirmed the important role of plants in IVCW-MFCs.     

Effects of SO2 and H2O on low-temperature NO conversion over F-V2O5-WO3/TiO2 catalysts

F-V_2 O_5-WO3/Ti02 catalysts were prepared by the impregnation method.As the content of F ions increased from 0.00 to 0.35 wt.%,the NO conversion of F-V_2 O_5-WO_3/TiO_2 catalysts initially increased and then decreased.The 0.2 F-V_2 O_5-WO_3/TiO_2 catalyst(0.2 wt.% F ion)exhibited the best denitration(De-NOx) performance,with more than 95% NO conversion in the temperature range 160-360℃,and 99.0% N2 selectivity between 110 and 280℃.The addition of an appropriate amount of F ions eroded the surface morphology of the catalyst and reduced its grain size,thus enhancing the NO conversion at low temperature as well as the sulfur and water resistance of the V_2 O_5-WO3/Ti02 catalyst.After selective catalytic reduction(SCR) reaction in a gas flow containing SO_2 and H_2 O,the number of NH3 adsorption sites,active component content,specific surface area and pore volume decreased to different degrees.Ammonium sulfate species deposited on the catalyst surface,which blocked part of the active sites and reduced the NO conversion performance of the catalyst.On-line thermal regeneration could not completely recover the catalyst activity,although it prolonged the cumulative life of the catalyst.In addition,a mechanism for the effects of S02 and H_2 O on catalyst NO conversion was proposed.     

Fast and efficient aqueous arsenic removal by functionalized MIL-100(Fe)/rGO/δ-MnO2 ternary composites: Adsorption performance and mechanism

Because of its significant toxicological effects on the environment and human health,arsenic(As) is a major global issue.In this study,an Fe-based metal-organic framework(MOF)(Materials of Institut Lavoisier:MIL-100(Fe)) which was impregnated with reduced graphene oxide(rGO) by using a simple hydrothermal method and coated with birnessitetype manganese oxide(δ-MnO_2) using the one-pot reaction process(MIL-100(Fe)/rGO/δ-MnO_2 nanocomposites) was synthesized and applied successfully in As removal.The removal efficiency was rapid,the equilibrium was achieved in 40 min and 120 min for As(Ⅲ) and As(Ⅴ),respectively,at a level of 5 mg/L.The maximum adsorption capacities of As(Ⅲ) and As(Ⅴ) at pH 2 were 192.67 mg/g and 162.07 mg/g,respectively.The adsorbent revealed high stability in pH range 2-9 and saturated adsorbent can be fully regenerated at least five runs.The adsorption process can be described by the pseudo-second-order kinetic model and Langmuir monolayer adsorption.The adsorption mechanisms consisted of electrostatic interaction,oxidation and inner sphere surface complexation.     

基于模糊PI控制的瞬态高温控制策略研究

目的通过地面环境试验研究气动热引起的瞬态高温环境对飞行器结构的影响。方法设计基于石英灯和离心机的瞬态高温-加速度复合试验系统,研究温度和加速度同步控制系统的结构和原理,分析无制冷装置环境下瞬态高温控制的难点,提出基于模糊PI控制器的温度控制策略。结果通过MATLAB仿真验证,设计的模糊PI控制器相比传统的PI控制器能够有效提高温度控制的动态性能与精确性。结论将基于模糊PI控制的瞬态加热控制策略应用于瞬态高温控制是可行、有效的,能为地面气动热模拟环境试验提供技术保障。     

弹射起飞和拦阻着陆冲击试验方法探讨

目的研究舰载机机载装备弹射起飞和拦阻着陆冲击试验的工程实现方法。方法从GJB 150.18A—2009、MIL-STD-810F/G及其提供的实测数据出发,分析弹射起飞和拦阻着陆冲击载荷特征,对几种常用试验方法的工程实现进行探讨和比较。结果弹射起飞和拦阻着陆冲击具有瞬态、交变、低频、大位移、高速度等特征。阻尼正弦波方法的实施路径比较明确,复杂波形再现方法实施路径比较模糊且复杂,超长脉冲半正弦波方法具有多种局限性,前两种方法的速度、位移都较大。结论阻尼正弦波方法较复杂波形再现方法更容易实现,但两者都需要大位移、高速度的专用冲击试验设备和适当的波形处理技术,超长脉冲半正弦波方法在前面两种标准推荐方法能够实现时,不建议使用。     

海水入侵指标对比分析与评价—以珠江口地下水含水层为例

合理评价海水入侵对于地下水含水层管理和居民生活健康有着重要的意义。当前海水入侵评价应用较广泛的为包含单一因子和基于统计分析的多种水化学方法。本文综合阐述了海水入侵评价中所用的水化学指标、原理与方法,包括在国内评价体系中少见的一些方法,并以珠江口地下水含水层为例,对比了各种指标的有效性,评价了各类方法的优缺点以及在实际应用中可能产生的问题。结果显示,以Cl-和TDS作为简单直接的单因子可以快速评价海水入侵,尤其对于大范围的海水入侵评价十分有效。而和Ca2+,HCO₃-,SO₄2-等变化有关的指标具有地域性,在研究区对海水入侵的指示并不是十分敏感。在复杂评价因子中,与离子交换有关SAR,BEX以及GQI_SWI与Cl-相关性较高,可以较好的指示海水入侵及水岩相互作用。经过分析,受到地下含水层和地表水道双重介质入侵的影响,研究区海水入侵范围在过去几个年代有向北推进的趋势。本研究对于高速经济发展下的沿海地下水含水层管理和评价有重要的指导意义。     

基于空间适宜性评价和人口承载力的贵溪市中心城区城市开发边界的划定

城市开发边界的划定能够合理引导城市空间的有序发展,控制其无序蔓延。论文以典型的资源型城市贵溪市为研究区,从空间适宜性评价、人口承载力、城市规模以及城市总体规划和土地利用总体规划（简称“两规”）衔接等方面探索有效的城市开发边界划定方法。空间适宜性评价综合考虑自然、空间可达性和生态条件,运用聚类分析法确定区域内不宜作为开发建设的生态底线区域和适宜建设开发区域的高低等级,以此确定城市开发边界的发展方向。通过灰色预测GM(1,1)模型预测研究区2020年人口总量,并基于土地资源和水资源承载力验证当地所能容纳的最大人口总量,同时确定城市规模和划定城市开发边界。以空间适宜性评价、人口承载力、城市规模预测、两规衔接和空间形态控制等方法倒逼缩减建设用地,从而划定城市发展的刚性和弹性增长边界。     

基于高分1号影像的森林植被信息提取

实时最新森林植被信息的提取是林业航空植保作业的必要前提。论文以安徽省蚌埠市为研究区域,探讨了基于高分1号卫星遥感数据在亚热带农林植被混合地区的森林植被信息提取。根据植被物候信息差异选择了提取森林植被信息的5个关键时期高分影像,采用分区决策树方法监测森林植被的空间分布和面积信息,并与未分区决策树法的提取结果进行比较。结果表明：采用分区决策树法和未分区决策树法对于大中尺度森林植被信息提取的总体精度均优于85%。但分区决策树森林植被提取总体精度达到90.72%,较未分区决策树法提高3.80%、4.65%,Kappa系数达到0.81,较未分区决策树法提高约0.07~0.10,结合植被物候信息的分区决策树森林植被提取法好于未分区决策树法,能够满足林业航空植保作业的精度需求。具有较高空间分辨率、宽覆盖、短重访周期的高分1号影像,对于大区域的林业航空植保当年最新森林植被信息的提取表现出较大的潜力。