进气相位对天然气转子发动机流场和燃烧过程的影响

转子发动机进气相位的改变不仅改变了进气效率,而且对缸内流场以及燃烧过程都产生很大的影响。基于FLUENT软件,通过编程实现了相应的三维动网格模型,添加了合适的湍流模型、燃烧模型、CHEMKIN化学反应机理,建立了基于化学反应动力学的三维动态数值模拟模型,并通过已有实验数据对比验证了模型的可靠性。在此基础上,对不同进气相位的周边进气天然气转子发动机的工作过程进行了数值模拟。计算结果表明:在进气持续期不变的情况下,随着进气开启角的提前,缸内涡团和旋流的强度、充量系数都不断增加,点火位置处的流场湍流度不断增加。火花塞点火后其附近的流场湍流度增加可以大幅度提高燃烧速率,但燃烧行程的整体燃烧速率不是一直随着进气开启角的提前而增大。相同进气持续期下,进气开启角为407°时,火焰传播速度最大,缸内示功图最好,同时NO生成量也最大。     

轻油基燃料蒸发性及其对火花点火发动机性能影响研究

轻油基燃料蒸发性与普通车用汽油相比差别较大，必然影响火花点火发动机及汽车的性能。作者通过理论分析和试验研究证实：轻油基燃料蒸发性好，可显著改善发动机低温起动、预热和加速性；在正常环境条件下，燃用2号轻油基燃料的发动机及车辆不发生气阻；1、3号轻油基燃料由于蒸气压较高，在夏季使用时采用电动燃油泵或性能良好的机械式燃油泵，可消除气阻。     

点火定时对缸内直喷汽油机燃烧及颗粒物排放的影响

采用燃烧分析仪和DMS500型快速颗粒取样分析仪,在一台缸内直喷汽油机上进行了点火定时对燃烧过程和颗粒物排放影响的试验研究.结果表明:随着点火定时的不断延迟,火焰发展期逐渐缩短,快速燃烧期逐渐增大,缸内压力峰值逐渐下降,瞬时放热率峰值和缸内最高燃烧温度均逐渐降低且后移,放热过程迟缓,膨胀行程缸内温度逐渐升高.缸内直喷汽油机排气颗粒物排放呈包括核态和积聚态颗粒物的双峰分布;随点火定时的推迟,核态和积聚态颗粒物峰值数密度均逐渐降低,颗粒物总数量浓度逐渐降低,积聚态颗粒物峰值粒径逐渐减小,而核态颗粒物峰值粒径受点火定时的影响较小.     

Soil‐moisture change caused by experimental extreme summer drought is similar to natural inter‐annual variation in a loamy sand in Central Europe

Intensification of weather extremes is currently emerging as one of the most important facets of climate change. Research frontiers are in analyzing (1) the consequences for the hydrological cycle and (2) the effects of multifactor scenarios on ecosystems. However, in all theoretical and experimental scenarios, challenges arise as to how precipitation regimes translate into variation in soil moisture. Here, we explore soil‐moisture response to experimental changes in the precipitation regime in Central Europe over a period of 5 y, particularly focusing on the effects of recurrent extreme weather events. Intraannual difference in weekly precipitation sums imposed by extreme‐drought or heavy‐rainfall manipulations clearly exceeded interannual variation in the ambient precipitation pattern during the growing season between 2005 and 2009. However, soil‐moisture variability in the experimental plots did not clearly reflect any altered patterns in response to the manipulated precipitation regime. Natural variation in soil moisture between years was similar to within‐season differences between manipulations. Strong differences in soil‐moisture dynamics during the growing season can, however, be generated by changing the temporal distribution of rainfall events while keeping the magnitude of the precipitation sum constant. Our findings confirm a common methodological dilemma in precipitation‐change experiments searching for a logical way to determine how precipitation change affects communities and ecosystems on relatively short time scales: Alteration of weather regimes according to extreme‐value statistics and future scenarios vs. systematic alteration of soil moisture. For Central Europe, our data suggest that other factors rather than the magnitude of rainfall exclusion or addition would appear to be decisive for ecosystem response to more extreme precipitation regimes. Response of soil moisture to frequency, return interval, and timing of events is a promising approach for further exploration. In addition, buffer capacity of the ecosystem under study has to be taken into account.     

水分管理对水稻生长期N2O排放的影响

通过温室盆栽试验研究相同施肥条件下,不同水分管理(持续淹水,分蘖肥后提前烤田、正常烤田和推迟烤田)对水稻生长期N2O排放的影响。结果表明,整个水稻生长期N2O的排放集中在施分蘖肥后的烤田期间及随后复水的5~8天内。在此期间各处理N2O排放量占季节排放总量的70.30%~94.26%;烤田前淹水及后期干湿交替阶段有少量N2O排放。对持续淹水处理而言,N2O的排放通量是正常烤田处理的12.10%。提前烤田、正常烤田和推迟烤田的平均N2O排放通量分别为188.56、158.48和147.84μg/(m2?h),随着烤田开始时间的推迟,N2O排放量逐渐减少。     

基于Spark框架XGBoost的林业文本并行分类方法研究

针对当前互联网+技术与林业的交叉融合,涌现出海量待挖掘的涉林文本,而林业文本分类的相关研究尚不成熟的问题,使用网络爬虫技术面向互联网采集涉林文本,基于丰富的语料重新构建分类标签,提出基于Spark计算框架的XGBoost并行化方法,对林业文本进行分类。经由交叉验证,构建的XGBoost并行分类算法准确率为0. 923 4,在各类别中最低F1为0. 860 4,最高为0. 998 4;其在2. 1万条、4. 2万条、8. 4万条数据集上的训练加速比分别为2. 13、3. 47、3. 82。结果表明,基于该标签设定的分类模型对现存互联网中涉林文本的适应性较好;Spark环境下实现的XGBoost并行化算法的准确率显著优于其他4种机器学习(朴素贝叶斯、GBDT决策树、BP神经网络和ELM神经网络算法)的并行化算法,算法执行效率远高于单机版本,且数据量越大,其加速比越高,能有效应对海量林业文本的实时、准确分类。     

基于Spark的WOA-BP水稻产量预测

目的 随着大数据技术和人工智能的快速发展,针对当前水稻产量预测模型精度低、预测区域范围过大、模型优化时间过长等问题,本文提出一种基于Spark的鲸鱼优化算法-反向传播神经网络(Whale optimization algorithm-backpropagation,WOA-BP)水稻产量预测方法。方法 本文以广东省西部地区的县/市/区水稻产量及气象数据作为研究对象,采用WOA对BP网络的权值和偏置值进行优化,并构建水稻产量预测模型,提升预测精度;此外,在Spark框架下,实现WOA-BP算法并行化,减少算法时间开销。结果 模型精度方面,通过对预测结果进行反归一化后比较,经WOA优化后的BP神经网络模型,平均绝对百分比误差 (Mean absolute percentage error) 从8.354%降至7.068%,平均绝对误差 (Mean absolute error) 从31.320 kg降至26.982 kg,均方根误差 (Root mean square error) 从41.008 kg降至33.546 kg;运行时间方面,3节点Spark集群比非Spark模式减少了11 742 s,减少44%的时间开销。结论 基于Spark的WOA-BP水稻产量预测方法,能够较好地预测出广东西部县/市/区的水稻产量,同时可以很好地反映气象因素对广东省西部地区水稻产量的影响情况,对研究广东西部县/市/区乃至整个广东的水稻产量情况具有一定的参考价值。     

小功率非道路用柴油机动力、经济及排放特性

为提高小功率非道路用柴油机的动力性、经济性、降低有害物排放,开展了柴油机缸内燃烧过程的试验研究。该文以非道路用N490直喷柴油机为样机,通过对燃料喷射系统参数优化,提高喷油压力和优化喷油特性;通过对配气定时调整,提高柴油机常用转速下的充量系数;通过对燃烧室结构的优化设计,改善缸内油气混合。柴油机经优化后,性能得到大幅提升,柴油机在标定工况、最大扭矩工况燃油消耗率分别下降了3.0%、2.43%,不透光烟度分别减少了14.6%、10%;整机的CO、NO_x+HC、颗粒(PM)与原机相比分别下降了34.7%、21.5%、34.9%,排放达到了非道路国三排放标准限值要求。通过对燃料喷射系统、进气系统与燃烧室的优化匹配,完全可以在降低有害物排放的同时,提高柴油机的经济性。试验研究可为提升非道路用柴油机性能并满足国三排放法规提供技术参考。     

Effect of time of application and amount of nitrogen fertilizer on seed yield and concentration of oil in canola (Brassica napus)

The seed (grain) yield increases (responses) and concentration of oil in seed responses of canola (Brassica napus L.) to applications of fertilizer nitrogen (N), as urea (46% N), was measured in eight field experiments in south-western Australia (SWA). Nitrogen was applied at five different times of application, either at sowing or at three to four weekly intervals until 12–16 weeks (0, 3, 6, 9, 12 or 0, 4, 8, 12, 16) after seedling emergence. Canola, sown in late May to early June, was grown on a range of soil types in different locations of SWA. The greater the amount of N applied and the closer N was applied to the sowing of the canola seed usually gave the largest seed yield increase at both higher rainfall sites (> 500 mm) and lower rainfall sites (<350 mm). Maximum seed yield of canola were reached within nine weeks after seedling emergence. The exception was for a sandy soil (Fluventic Lithic Xerochrept; Brown Tenosol) at Narrogin where applications of N at six, nine, or 12 weeks after emergence gave higher yields compared to N applied earlier mainly due to N leaching in June and July.

The amount of N required for 90% of maximum seed (N_90%Y) yield ranged from five to 58 kg N ha^?1 with the amount depended on location and growing season. For six of the eight sites the higher amounts of N for N_90%Y were required at sowing and three weeks after emergence. Similarly, N use efficiency (NUE, kg grain produced kg N applied^?1) tended was highest for either the N applied at sowing or within three to four weeks after emergence of seedlings. The exception was for a sandy soil at Narrogin where applications of N at six, nine, or 12 weeks after emergence gave higher NUE compared to N applied earlier. N use efficiency decreased as the amount of N increased for all times of N application. Generally, the amount of N applied decreased the oil concentration of canola seed at each time of application. However, the effect of the time of application of N fertilizer on the decrease in oil concentration of canola seed was largest with the highest N level applied at 12 or 16 weeks after seedling emergence. The percentage the oil concentrations decreased as the amount of N applied increased varied with location and growing season. Further research work is required to elucidate the interaction between the growing season, possible rainfall and temperature, and the effects of N on grain yield and oil concentration in seed.     

Analysis and Research on VVT System Control Strategy for Motorcycle Engine

An original variable valve timing mechanism with two-mode intake timing parameters is developed according to the structural characteristics of small-medium displacement motorcycle engine. Based on cycle simulation, a cycle simulation model of VVT mechanism for motorcycle engine is created. The control strategy of variable valve timing and lifting running-model are carried out at idle speed low negative load, middle negative load, low speed high negative load and high speed high negative load. It has been proved that the performance of JH125 motorcycle engine can be effectively improved by loading the new type VVT system.