浅议生态环境监测机构的风险控制

现阶段,为了探索创新性的发展途径,生态环境监测机构应始终增强风险控制意识。应加强风险管理,立足于风险控制,以求不断健康发展。     

一种飞行计划航段自动激活方法

飞机从一个航段飞向下一个航段时,需要对下个航段是否激活进行判断.提出了一种新的飞行计划航段自动激活方法,通过获取飞机当前位置及其磁差,并结合临近的3个航路点对应位置的磁差经纬度,分别计算3个航路点与飞机位置对应的磁航向角,通过磁航向角之间做差进行比较,进而判断是否被激活.算法推导完成后,通过基于模型的系统工程工具SCADE SUITE进行了建模,并结合杰普逊航图进行了实际数据采样与仿真.仿真结果表明:该航段自动激活方法能够有效稳定地激活下一航段,比传统方法计算简便,方法实用有效.已在实际飞行中成功应用,达到了预期效果.     

Effect of phosphogypsum on saline-alkalinity and aggregate stability of bauxite residue

A column experiment was conducted to investigate the effect of phosphogypsum (PG) on the saline- alkalinity, and aggregate stability of bauxite residue. Results showed that: with increasing leaching time, the concentrations of saline-alkali ions decreased while the concentration increased in bauxite residue leachate; compared with CK (control group) treatment, pH, electric conductivity (EC), exchangeable sodium percentage (ESP), sodium absorption ratio (SAR), and exchangeable Na⁺ content of bauxite residue were reduced following PG treatment; average particle sizes in aggregates following CK and PG treatments were determined to be 155 and 193 nm, respectively. SR-μCT test results also confirmed that bauxite residue following PG treatment acquired larger aggregates and larger pore diameter. These results indicate that the PG treatment could significantly modulate the saline-alkalinity, and simultaneously enhance aggregate stability of bauxite residue, which provides a facile approach to reclaim bauxite residue disposal areas.     

Generating Questions Based on Semi-Automated and End-to-End Neural Network

With the emergence of large-scale knowledge base, how to use triple information to generate natural questions is a key technology in question answering systems. The traditional way of generating questions require a lot of manual intervention and produce lots of noise. To solve these problems, we propose a joint model based on semi-automated model and End-to-End neural network to automatically generate questions. The semi-automated model can generate question templates and real questions combining the knowledge base and center graph. The End-to-End neural network directly sends the knowledge base and real questions to BiLSTM network. Meanwhile, the attention mechanism is utilized in the decoding layer, which makes the triples and generated questions more relevant. Finally, the experimental results on SimpleQuestions demonstrate the effectiveness of the proposed approach.     

碳脱氧在钢包顶渣改质中的应用

 为了进一步降低夹杂物缺陷并提高产品质量,基于碳脱氧进行了钢包顶渣改质的研究。冷轧产品的生产工艺为铁水预处理→转炉→RH精炼→连铸,为减少钢中夹杂物质量分数,需要进行钢包顶渣改质,同时降低钢包顶渣TFe质量分数。采用粒碳部分替代铝渣球的方法进行基于碳脱氧工艺的钢包顶渣改质,试验结果表明,顶渣改质效果良好,在顶渣TFe质量分数、中间包钢水游离氧明显降低的同时铸坯中Al₂O₃夹杂物得到优化;“30 kg粒渣+铝渣球”工艺降低生产成本5.16元/t(钢)。     

基于GFU和分层LSTM的组群行为识别研究方法

提出一种以"关键人物"为核心，使用门控融合单元（GFU，Gated Fusion Unit）进行特征融合的组群行为识别框架，旨在解决两个问题：①组群行为信息冗余，重点关注关键人物行为特征，忽略无关人员对组群行为的影响；②组群内部交互行为复杂，使用GFU有效融合以关键人物为核心的交互特征，再通过LSTM时序建模成为表征能力更强的组群特征.最终，通过softmax分类器进行组群行为类别分类.该算法在排球数据集上取得了86.7%的平均识别率.     

Ca-Mg-Al复合氧化物催化尿素与1,2-丙二醇制碳酸丙烯酯

采用共沉淀法制备了系列Ca-Mg-Al复合氧化物催化剂，通过对沉淀剂配比、沉淀液pH值及焙烧温度等制备条件的考察，得到以Na₂CO₃为沉淀剂、pH=9.5、850℃下焙烧4h制备的Ca-Mg-Al催化活性最高。在 n(PG)∶n(urea)=1.5∶1、反应温度为145℃、绝压20kPa、反应时间4h、催化剂用量为尿素质量的5%时，碳酸丙烯酯收率达到84.6%。采用XRF、XRD、NH₃-TPD、SEM及BET对催化剂的组成、晶型及酸性进行了表征，发现随着沉淀剂中Na₂CO₃的含量增加，催化剂中CaO∶MgO的比例增大，碳酸丙烯酯的收率亦升高；经850℃焙烧后，催化剂中存在CaO和MgO两种活性中心，起协同催化作用；随着焙烧温度由700℃升高到850℃，NH₃-TPD脱附曲线向低温方向偏移，且强酸中心NH₃脱附峰面积比由81.14%明显下降为0，中强酸中心NH₃脱附峰面积比由0增加到78.07%，而碳酸丙烯酯收率由68%增加到84.6%，这表明催化剂强酸性位的减少是催化活性增加的主要原因。     

基于分段加权的反向稀疏跟踪算法研究

为提高稀疏表示跟踪模型性能，提出一种分段加权的反向稀疏跟踪算法，将跟踪问题转化为在贝叶斯框架下寻找概率最高的候选对象问题，构造不同的分段权重函数来分别度量候选目标与正负模板的判别特征系数。通过池化来降低跟踪结果的不确定性干扰，选择正负模板加权系数差值最大的候选表示作为跟踪结果。实验表明，在光照变化、遮挡、快速运动、运动模糊情况下，所提出的算法可以确保跟踪结果的准确性和鲁棒性。     

Spreading and Curing Behaviors of a Thermosetting Droplet-Silicone on a Heated Surface

Thermosetting materials are widely used as encapsulation in the electrical packaging to protect the core electronic components from external force, moisture, dust, and other factors. However, the spreading and curing behaviors of such kind of fluid on a heated surface have been rarely explored. In this study, we experimentally and numerically investigated the spreading and curing behaviors of the silicone(OE6550 A/B, which is widely used in the light-emitting diode packaging) droplet with diameter of ～2.2 mm on a heated surface with temperature ranging from 25 ℃ to 250 ℃. For the experiments, we established a setup with high-speed camera and heating unit to capture the fast spreading process of the silicone droplet on the heated surface. For the numerical simulation, we built a viscosity model of the silicone by using the Kiuna's model and combined the viscosity model with the Volume of Fluid(VOF) model by the User Defined Function(UDF) method. The results show that the surface temperature significantly affected the spreading behaviors of the silicone droplet since it determines the temperature and viscosity distribution inside the droplet. For surface temperature varied from 25 ℃ to 250 ℃, the final contact radius changed from ～2.95 mm to ～1.78 mm and the total spreading time changed from ～511 s to ～0.15 s. By further analyzing the viscosity evolution of the droplet, we found that the decreasing of the total spreading time was caused by the decrease of the viscosity under high surface temperature at initial spreading stage, while the reduction of the final contact radius was caused by the curing of the precursor film. This study supplies a strategy to tuning the spreading and curing behavior of silicone by imposing high surface temperature, which is of great importance to the electronic packaging.