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.     

大纵坡城市道路透水边带截留能力试验

道路是城市重要的交通纽带,一旦发生积水内涝,将导致城市交通瘫痪,甚至危及生命财产安全,特别是山地城市道路普遍存在纵坡大、雨水口截留效率低等突出问题,发生积水内涝和马路洪水的风险较高。针对上述问题,提出了利用透水边带提高大纵坡城市道路雨水径流截留效率的方法。在实验室搭建了城市单车道物理试验模型(比例1∶1),采用人工模拟降雨方法,系统研究了不同透水面积比、不同重现期降雨条件下透水边带对雨水径流的截留效率,并与传统道路雨水口截留能力进行了比较。试验结果表明:相同透水面积比条件下,雨水径流截留能力随着重现期的增大而减小;相同重现期条件下,雨水径流截留能力随着透水面积比的增加而增大,在重现期P=5 a时,透水边带面积比从12.5%增加到50%,雨水径流截留率从72.3%增加到79.3%。与传统雨水口截留能力相比,增加透水边带后雨水径流截留能力可提高30%左右。因此,大纵坡城市道路在不影响交通安全的条件下,可根据道路空间布局特征,通过设置透水边带来提高雨水径流的截流效率。     

对低影响开发与海绵城市的再认识

海绵城市建设是在继承我国古代先贤智慧和参考国外经验,系统总结我国雨洪管理领域长期研究和实践经验的基础上,结合我国城市水系统实际问题提出的城市发展方式,其核心是构建基于绿灰结合的现代城市雨洪控制系统,通过"渗、滞、蓄、净、用、排"综合措施,实现"治涝"与"治黑"等多重目标。低影响开发是海绵城市建设的重要指导思想,也是海绵城市核心技术体系的重要组成部分。正确认识低影响开发与海绵城市的内涵与联系,对于进一步在全国范围内落实低影响开发建设模式,科学推进海绵城市建设具有重要意义。     

基于洪涝风险分析的纽约市城市绿地规划设计及其对中国的启示

1968年美国开始推行洪涝保险计划,不断完善对洪涝风险的研究,并逐步形成了一套相对完善的洪泛区管理体系,而洪涝风险分析在洪涝保险、城市规划、土地开发、应急管理等领域广泛应用。纽约市在经历了多次飓风侵害,尤其是2012年飓风桑迪（Sandy）之后,意识到城市绿地在极端暴雨事件时对雨洪调蓄的重要作用,经过持续的研究实践,提出了基于洪涝风险分析的城市绿地规划设计要求。基于总结美国纽约市在飓风桑迪影响下对洪涝风险图的调整,及其对城市绿地规划设计相关要求,提出其对我国洪涝风险管控及城市绿地规划设计的启示。     

基于“BIM+GIS”技术的建筑垃圾精准管控信息管理平台研究初探

近年来,我国建筑垃圾产量不断增加,全国600多座大中城市中有70%被垃圾所包围,其中建筑垃圾贡献率达到40%,而相比于部分发达国家,我国尚缺少针对于建筑垃圾的全过程实时监测与智能管控手段。将"BIM+GIS"技术引入建筑垃圾监管领域,并基于此技术建立建筑垃圾实时监测与智能管控信息管理平台,可实现对建筑垃圾的精准管控。     

基于在线监测+模型的海绵城市建设径流水量控制效果研究

以青岛市海绵城市典型项目区为研究区,通过收集、整理和概化汇水区资料,构建了研究区降雨径流模型——SWMM模型。利用2018年雨季的降雨-径流实时监测数据,进行了模型关键参数的率定和模型验证,研究了海绵城市建设在削减径流量、增加入渗与滞蓄雨量方面的定量效果。结果显示:海绵改造后,不同降雨强度下形成径流的雨量占总雨量的平均比例由56%下降至29%;下渗雨水的平均比例由40%增加至60%,滞蓄深度由0.63 mm增加至5 mm。海绵改造前后研究区场降雨径流总量控制率变化显著,当以1~10年一遇的短历时（120 min）设计降雨作为模型边界条件进行模拟时,海绵化改造后场次降雨径流总量控制率提高了26%~34%,认为研究区基本达到了海绵城市关于水量控制的建设要求。     

基于Stacking的地面PM2.5浓度估算

为了解决地面PM_2.5监测网络在空间和时间覆盖受限的问题,提出了基于宽时空覆盖的卫星气溶胶光学厚度AOD,利用Stacking方法建立地面PM_2.5浓度估算模型,将AOD、PM_2.5和各气象参数以及与PM_2.5排放有关的数据进行训练,使用改进网格搜索对模型超参数进行优化,通过对多重共线性分析,建立基于Stacking的最优PM_2.5浓度估算模型。选取2016-01-01-2016-12-31的数据作为实验对象,结果表明:相比于随机森林、GBRT和XGBoost 3种模型,使用岭回归作为元学习器的Stacking模型性能更优,可见Stacking适用于大范围地理区域的大气污染监测。     

雨水管道沉积物中典型无机氮的干期粒径分布

以南京江北新区分流制雨水管道沉积物为研究对象,考察不同粒径沉积物中铵态氮（NH₃-N）和硝态氮（NO₃^--N）干期分布特征,分析其随干期长度的变化关系,并探讨沉积物理化性质及微生物菌群结构对NH₃-N和NO₃^--N干期分布的影响.结果表明：粒径≤0.075mm的沉积物中NH₃-N占比最高（交通区30.8%,商业区36.7%）;交通区粒径≤0.075mm的沉积物中NO₃^--N占比最高（33.0%）,而商业区粒径0.075~0.15mm沉积物中NO₃^--N占比最高（34.4%）;干期长度与交通区0.075~0.15mm粒径段的沉积物中NO₃^--N含量及商业区粒径≥0.3mm的沉积物中NH₃-N含量之间的相关性均最显著;雨水管道沉积物中NH₃-N和NO₃^--N的干期分布与O-H和N-H等官能团、表面极性和亲水性、微观形貌等有一定关联;交通区雨水管道沉积物中Gemmobacter等反硝化优势菌种（相对丰度总和为20.15%）对NO₃^--N干期分布影响更显著.     

Acceleration of floc-water separation and floc reduction with magnetic nanoparticles during demulsification of complex waste cutting emulsions

Waste cutting emulsions are difficult to treat efficiently owing to their complex composition and stable emulsified structure. As an important treatment method for emulsions, chemical demulsification is faced with challenges such as low flocs–water separation rates and high sludge production. Hence, in this study, Fe₃O₄ magnetic nanoparticles (MNPs) were used to enhance chemical demulsification performance for treating waste cutting emulsions under a magnetic field. The addition of MNPs significantly decreased the time required to attain sludge–water separation and sludge compression equilibrium, from 210 to 20 min. In addition, the volume percentage of sludge produced at the equilibrium state was reduced from 45% to 10%. This excellent flocculation–separation performance was stable over a pH range of 3–11. The magnetization of the flocculants and oil droplets to form a flocculant–MNP–oil droplet composite, and the magnetic transfer of the composite were two key processes that enhanced the separation of cutting emulsions. Specifically, the interactions among MNPs, flocculants, and oil droplets were important in the magnetization process, which was controlled by the structures and properties of the three components. Under the magnetic field, the magnetized flocculant–MNP–oil droplet composites were considerably accelerated and separated from water, and the sludge was simultaneously compressed. Thus, this study expands the applicability of magnetic separation techniques in the treatment of complex waste cutting emulsions.     

Investigation to meet China II emission legislation for marine diesel engine with diesel methanol compound combustion technology

In order to reduce the pollutant emission and alleviate the pressure of petroleum resources shortage and greenhouse gas emission at the same time, the use of clean and renewable alternative fuel for marine engines is a promising option. In this study, a marine diesel engine, which was modified to run in diesel methanol compound combustion (DMCC) mode, was investigated. After the diesel injection parameters were calibrated, and combined with a sample after-treatment device DOC (diesel oxidation catalyst), the engine could meet the requirements of China II legislation. The overall MSP (methanol substitute percent) reached 54.1%. The value of each pollutant emission was much lower than that in China II emission legislation, and there was almost no methanol and formaldehyde emissions. When methanol was injected into the inlet manifold, the intake air temperature decreased a lot, as well as the exhaust gas temperature, which were beneficial to increase engine thermal efficiency and improve engine room environment. Compared with the engine running in pure diesel mode, when the engine ran in diesel/methanol dual fuel mode, the combustion phase was advanced, and the combustion duration became shorter. Therefore, the engine thermal efficiency increased, and fuel consumption decreased significantly.