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.     

Kitchen waste valorization through a mild-temperature pretreatment to enhance biogas production and fermentability: Kinetics study in mesophilic and thermophilic regimen

Biowaste valorization through anaerobic digestion is an attractive option to achieve both climate protection goals and renewable energy production. In this paper, a complete set of batch trials was carried out on kitchen waste to investigate the effects of mild thermal pretreatment, temperature regimen and substrate/inoculum ratio. Thermal pretreatment was effective in the solubilisation of macromolecular fractions, particularly carbohydrates. The ability of the theoretical methodologies in estimating hydrogen and methane yields of complex substrates was evaluated by comparing the experimental results with the theoretical values. Despite the single batch configuration, a significant initial hydrogen production was observed, prior to methane yield. Main pretreatment effect was the gain in hydrogen production; the extent was highly variable according to the other parameters values. High hydrogen yields, up to 113 mL H2/g VSfed, were related to the prompt transformation of soluble sugars. Thermophilic regimen resulted, as expected, in faster digestions (up to 78 mL CH₄/gVS/day) and sorted out pH inhibition. The relatively low methane yields (342–398 mL CH₄/g VS_fed) were the result of the consistent lignocellulosic content and low lipid content. Thermal pretreatment proved to be a promising option for the enhancement of hydrogen production in food waste dark fermentation.     

Removal of high concentrated ammonia nitrogen from landfill leachate by landfilled waste layer

The landfill of municipal solid waste(MSW) could be regarded as denitrification reactor and involved in ammonia nitrogen biological removal process. In this research, the process was applied to municipal solid waste collected in Shanghai, China, which was characterized by high food waste content. The NH4^ removal efficiency in the system of SBR nitrifying reactor followed by fresh and matured landfilled waste layer in series was studied. In the nitrifying reactor, above 90% of NH4^ in leachate was oxidized to NO2^- and NO3^- . Then high concentrated NO2^- and NO3^- were removed in the way of denitrification process in fresh landfilled waste layer. At the same time, degradation of fresh landfilled waste was accelerated. Up to the day 120, 136.5 gC/(kg dry waste) and 17.9 gN/(kg dry waste) were converted from waste layer. It accounted for 50.15 % and 86.89 % of the total carbon and nitrogen content of preliminary fresh waste, which was 4.42 times and 5.17 times higher than that of reference column respectively. After filtering through matured landfilled waste, BOD5 concentration in leachate dropped to below 100 mg/L, which would not affect following nitrification adversely. Because the matured landfilled waste acted as a well methanogenic reactor, 23% of carbon produced accumulatively from fresh landfilled waste degradation was converted into CH4.     

室内装修后空气中微量氨检测方法的研究

本文简述了室内有害气体氨气污染对人类健康的危害 ,详细介绍了日前国内外测量室内空气中氨含量所应用的不同方法 ,研究讨论了各种方法的最低检测浓度 ,在具体检测中各种不同方法的干扰离子以及干扰离子所允许的最高浓度 ,比较各种不同方法的优劣     

选择性离子交换法脱除水中NH3-N

采用储量较大的黑龙江省嫩江天然斜发沸石,对其脱除NH_3-N的能力、预处理及洗脱再生方法、交换平衡、交换动力学等进行系统的研究,并对沸石床层的交换过程进行数学模型计算。     

燃煤锅炉氨法烟气脱硫

通过氨法脱硫工艺在75t燃煤锅炉烟气脱硫工程中的运用,阐述了氨法脱硫的原理及工业条件,并对运行指标及处理效果进行分析。     

石墨炉原子吸收光谱法测定轻质油中痕量砷时样品前处理方法的探讨

研究了轻质油中痕量砷的几种样品处理方法。实验表明，石墨炉原子吸收光谱法测定时，用碘-甲苯和H2O2-HNO3相结合处理轻质油样品较为合适，萃取效率可达90％。并对影响萃取测定的其它因素作了初步探讨。     

复合诱变菌处理氨氮废水

采用诱变技术对微生物菌种进行了诱变处理，分别研究了紫外线诱变菌、硫酸二乙酯（DES）诱变菌和复合诱变菌对含氨废水的降解性能。结果表明，复合诱变后获得的变异菌株Z5对含氨废水中NH4^ -N的去除率比单独采用紫外线或DES诱变后获得的菌株有明显的提高。混合菌种对化肥行业废水中COD的NH4^ -N的去除率可以达到95%～98%，出水能达到行业一级排放标准。     

Seed yeast cultivation for salad oil manufacturing wastewater treatment

The mixture of five yeast strains obtained from soil could remove about 85% TOC of oil-rich wastewater in batch test.While the highest MLSS was obtained at an N:C of 1:5,the oil removal decreased with the increase of N:C during yeast sludge cultivation.Ammonium chloride was the best nitrogen source for yeast cultivation from the viewpoint of yeast growth and oil utilizxation.An ammonia concentration of over 1300 mg/L led to mass death of yeast at a pH of 5.The ammonia concentration should be controlled at a level of 1000 mg/L or lower.     

合成氨装置重点部位的安全管理及防范措施

对安庆分公司化肥厂合成氨装置 4年以来的运行及安全管理情况作了综述 . 对装置的重点 监督部位加强技术型管理 , 并实行了一系列行之有效的安全防范措施 , 使之达到了国内同类 型装置先进水平 .