Effect of solvent on the crystal morphology of royal demolition explosive

Important crystal faces that dominate the crystal morphology of royal demolition explosive (RDX) in vacuum were analyzed with the attachment energy (AE) method. Molecular dynamics (MD) simulations were used to calculate the interaction energies between these crystal faces and different solvent molecules for an attachment energy correction. Growth habits in the presence of different solvents were generated. The results showed that some crystal faces in solutions became morphologically more important than that in vacuum while others became less important. Thus, crystal shape and surface property changed a lot with the variation of crystal faces. The results from calculation were in agreement with those from the re-crystallization experiment, which indicated that cyclohexanone (CH) was a promising solvent to modify the crystal morphology of RDX for obtaining products with regular shape and high purity, while butyrolactone (BL) played a great role in improving the surface electrostatic property of RDX.     

基于高斯混合模型的一种工业供应链销售预测方法

针对定性工业供应链销售预测方法中存在的精度低、非智能化且无法处理复杂样本等问题,提出一种基于高斯混合模型的工业供应链销售预测方法.分析真实销售数据的特征信息及特征间的相关程度;在不同聚类簇数和特征数的组合下利用训练数据集迭代计算模型参数;对测试数据集进行销售预测,通过比较预测结果的准确率、召回率和F值三个指标确定可实现较好结果的特征值与聚类簇数目.实验结果表明,与人工神经网络模型和卷积神经网络模型相比,该方法在三个指标方面均表现出更好性能,可较好地预测销售成交情况.     

目标检测中语义约束检查算法的研究与实现

基本的目标检测任务是在图像中识别目标,并标注目标的类别和位置信息。但是,很多应用中的目标检测任务常常带有语义约束,典型的包括单类别目标的数量约束和多个目标之间的空间位置约束。如在基于视频的生产安全监控系统中,目标检测不仅要识别和标定安全防护装备,还要检测这些安全防护装备是否被规范穿戴。提出了一种目标检测中语义约束检查算法,定义一种语义约束的模型,然后对图像进行带有语义信息的目标检测,最终对目标检测结果与语义约束进行一致性判定。以电力施工防护装备检查的实际需求和现场安监视频为例,验证了所提出的目标检测中语义约束检查算法的有效性。     

咸宁深层地热资源潜力浅析

阐述咸宁深层地热形成条件、资源类型、可能存在的热储和载体特征、开发利用条件等,提出开发咸宁深层地热的思路,积极引进吸收国内外先进技术和设备,早日实现咸宁深层地热开发利用方面的突破。